CN114080622A

CN114080622A - System and method for sizing image-based disposable articles

Info

Publication number: CN114080622A
Application number: CN202080049460.7A
Authority: CN
Inventors: S.K.斯坦利; A.P.拉帕奇; A.J.索尔; A.E.昂格
Original assignee: Procter and Gamble Co
Current assignee: Procter and Gamble Co
Priority date: 2019-07-19
Filing date: 2020-07-13
Publication date: 2022-02-22
Also published as: WO2021016634A1

Abstract

The present invention provides systems and methods for recommending the size of a disposable article to be worn by a subject. The image of the subject is analyzed to determine a fit parameter of the subject. These fit parameters are applied to various disposable article sizing models.

Description

System and method for sizing image-based disposable articles

Technical Field

The present disclosure relates to systems and methods for determining a recommended size of a disposable article to be worn by a subject, and more particularly, to systems and methods for determining a recommended size of a disposable article to be worn by a subject based on processing of an image of the subject.

Background

Absorbent articles such as diapers, training pants, and the like are manufactured in a variety of sizes and configurations. To ensure that the absorbent article is comfortable and performs properly, it is important to wear the proper size. With multiple product lines available for purchase, and with each product line typically having multiple sizing options, determining which product line and which size are appropriate for a particular wearer can be challenging. In addition, disposable article sizes are often recommended based on the weight of the wearer rather than the physical dimensions of the wearer. However, the weight of the wearer may be a poor predictor of the actual physical dimensions of the wearer, and thus selecting a product size based on the weight of the wearer may result in improper fit. Further, while knowledge of the physical dimensions of the wearer is relevant to proper disposable article fit, the physical dimensions of the wearer are typically not known and can be difficult to measure manually. Accordingly, there is a need for systems and methods for recommending absorbent articles based on physical attributes of the wearer.

Disclosure of Invention

In one form, a computer-based method includes: storing, by a disposable article recommendation computing system, a plurality of disposable article sizing models in a data store, the plurality of disposable article sizing models corresponding to a respective plurality of pre-manufactured disposable articles available for purchase. The method further comprises the following steps: receiving, by the disposable article recommendation computing system, an image collected by at least one camera, wherein the image comprises a representation of a subject, and the subject is a consumer of a pre-made disposable article. The method further comprises the following steps: determining, by the disposable article recommendation computing system, a scale of the image that correlates a size of the representation of the subject in the image to a physical size of the subject. The method further comprises the following steps: determining, by the disposable article recommendation computing system, a physical attribute of the representation of the subject in the image. The method further comprises the following steps: determining, by the disposable article recommendation computing system, a plurality of fit parameters for the subject based on the scale of the image and the physical attribute of the representation of the subject. The method further comprises the following steps: applying, by the disposable article recommendation computing system, the plurality of fit parameters to one or more of the plurality of disposable article sizing models. The method further comprises the following steps: determining, by the disposable article recommendation computing system, a recommended pre-made disposable article for the subject based on application of the plurality of fit parameters to one or more of the plurality of disposable article sizing models, wherein the recommended pre-made disposable article is selected from a plurality of pre-made disposable articles available for purchase. The method further comprises the following steps: providing, by the disposable recommendation computing system, an indication of a recommended pre-made disposable for the subject.

In another form, a computer-based system includes: a data store, wherein a plurality of disposable article sizing models are stored by the data store, the plurality of disposable article sizing models corresponding to pre-made disposable articles of respective sizes available for purchase. The system further comprises: a disposable article recommendation computing system includes a computer readable medium having computer executable instructions stored thereon. The computer-executable instructions are configured to direct one or more computer processors to: receiving an image of a subject collected by a remote mobile computing device; determining a scale of the image; and processing the image to determine a physical attribute of the subject. Based on the scale of the image and the physical attributes of the subject, a plurality of fit parameters of the subject are determined. The computer-executable instructions are further configured to instruct the one or more computer processors to compare the plurality of fit parameters to one or more of a plurality of disposable article sizing models. Determining a recommended pre-made disposable article for the subject based on the comparison of the plurality of fit parameters to one or more disposable article sizing models of the plurality of disposable article sizing models. The computer-executable instructions are configured to instruct the one or more computer processors to send an indication of a recommendation for a pre-manufactured disposable article to a remote mobile computing device.

In another form, a computer-based method includes: a plurality of disposable article sizing models are stored, the plurality of disposable article sizing models corresponding to a respective plurality of pre-made disposable articles available for purchase. The method further comprises the following steps: an image collected by at least one camera is received, wherein the image includes a representation of a subject, and the subject is a consumer of a pre-made disposable article. The method further comprises the following steps: a scale of the image is determined that relates a size of the representation of the subject in the image to a physical size of the subject. The method further comprises the following steps: a physical property of a representation of a subject in an image is determined by image processing. The method further comprises the following steps: a plurality of fit parameters of the subject are determined based on the scale of the image and the physical attribute of the representation of the subject. The method further comprises the following steps: determining, by the disposable article recommendation computing system, a recommended pre-made disposable article for the subject based on application of the plurality of fit parameters to one or more of the plurality of disposable article sizing models, wherein the recommended pre-made disposable article is selected from a plurality of pre-made disposable articles available for purchase. The method further comprises the following steps: providing, by the disposable recommendation computing system, an indication of a recommended pre-made disposable for the subject.

Drawings

The above-mentioned and other features and advantages of this disclosure, and the manner of attaining them, will become more apparent and the disclosure itself will be better understood by reference to the following description of non-limiting embodiments of the disclosure taken in conjunction with the accompanying drawings, wherein:

FIG. 1 depicts a user interacting with an exemplary disposable article recommendation computing system.

FIG. 2 depicts a user interacting with another exemplary disposable article recommendation computing system.

Fig. 3A-3C depict exemplary scale objects and subjects placed within the field of view of an image capture device.

Fig. 4-7 depict various exemplary operating arrangements for a disposable article recommendation computing system.

FIG. 8 depicts an exemplary process that may be performed by the disposable article recommendation computing system.

Figure 9 depicts the determination of fit parameters for a subject using a scaled object.

Figure 10 depicts the determination of fit parameters for a subject using foot size determination scaling.

Figure 11 depicts the determination of fit parameters for a subject using head circumference determination scale.

FIG. 12 depicts a simplified sizing model of a disposable array.

FIG. 13 depicts a series of simplified interfaces for image collection and recommendation display.

FIG. 14 depicts a series of simplified interfaces for image collection using a scaled object graphical guide area.

Fig. 15-16 depict simplified interfaces displaying exemplary consumption predictions.

Fig. 17-19 depict simplified interfaces displaying exemplary purchase paths.

FIG. 20 depicts a series of simplified interfaces for collecting various inputs from a user and displaying recommendations.

FIG. 21 is a flow chart of an exemplary method of recommending a pre-made disposable article for a subject.

Detailed Description

The present disclosure relates to systems and methods for recommending the size of a disposable article based on processing of an image to determine physical attributes of an intended wearer. Various non-limiting embodiments of the present disclosure will now be described to provide an overall understanding of the principles of the function, design, and operation of the manufacturing systems and methods. One or more examples of these non-limiting embodiments are illustrated in the accompanying drawings. Those of ordinary skill in the art will understand that the systems and methods described herein and illustrated in the accompanying drawings are non-limiting exemplary embodiments and that the scope of the various non-limiting embodiments of the present disclosure is defined solely by the claims. The features illustrated or described in connection with one non-limiting embodiment may be combined with the features of other non-limiting embodiments. Such modifications and variations are intended to be included within the scope of the present disclosure.

As used herein, the term "absorbent article" refers to disposable devices that are placed against or in proximity to the body of a wearer to absorb and contain the various exudates discharged from the body, such as infant diapers, child diapers, adult diapers, pant diapers, training pants, feminine hygiene articles, and the like. Typically, these articles comprise a topsheet, a backsheet, an absorbent core, an acquisition system (which may be referred to as a liquid management system and may be composed of one or several layers), and generally other components, wherein the absorbent core is generally positioned at least partially between the backsheet and the acquisition system or between the topsheet and the backsheet. The absorbent article of the present disclosure will be further explained in the form of a taped diaper in the following description and the accompanying drawings. However, the description should not be taken as limiting the scope of the claims. Rather, the present disclosure applies to any suitable form of absorbent article (e.g., training pants, feminine hygiene products, adult incontinence products, etc.). For example, the systems and methods described herein are applicable to a range of different absorbent article types, such as disposable, semi-durable, single-use, multi-part, cloth, pant, pull-on, or insertion type absorbent articles and products. Absorbent articles according to the present disclosure may be pre-manufactured such that the absorbent articles are manufactured in a predetermined size configured to be worn by a wearer having certain physical attributes. In some embodiments, absorbent articles according to the present disclosure are at least partially customizable, such as certain aspects configurable based on physical attributes of the intended wearer. By way of example, the leg hoop size of an absorbent article to be worn by a particular wearer may be sized to provide better fit for that particular wearer.

In some configurations, the systems and methods described herein may receive one or more digital images of a subject and determine physical attributes of the subject through various image processing techniques. The subject may be, for example, an infant, a baby, a toddler, or other wearer of an absorbent article. The particular physical attributes determined by the systems and methods of the present invention may vary based on the implementation, but in some configurations, image analysis is performed to determine various fit parameters. Examples of fit parameters include an estimated waist circumference of the subject, an estimated thigh circumference of the subject, and an estimated rise measurement of the subject as measured from the navel to the lower back. Fit parameters can be applied to various sizing models of absorbent articles to assess which product or products will fit the subject. The subject's absorbent article recommendations may then be provided. The recommendation may identify, for example, any of a product size, a product type, and a product family. In some configurations, additional information about the subject in addition to the image may be used to determine the recommendation. Non-limiting examples of additional information about a subject that may be utilized by the systems and methods described herein include, but are not limited to, age of the subject, gestational age of the subject, geographic location of the subject, and developmental indicators completion. Exemplary developmental indicators may include, but are not limited to, crawling, climbing on furniture, walking, starting toilet training, and the like. Additionally or alternatively, dimensional information about the subject, such as the subject's height, weight, head circumference, etc., may be provided by the user, which may be utilized by the system in making absorbent article recommendations. Additionally or alternatively, fit assessments or performance feedback regarding currently used products, as well as other usage-based parameters such as the number of bowel diapers per day or the number of wet diapers per day, may also be provided as inputs. Additionally or alternatively, absorbent article preferences such as preferences for more natural products, preferences for products suitable for more active children, preferences for high absorbency or overnight products, and the like may be provided by and considered by the user.

Other user-provided information may include, for example, whether the wearer is wearing clothing or a diaper only, or the information may be determined by, for example, an algorithm. In any of the above examples, data associated with the user profile or other user-provided information may be used to determine the recommendation. Additionally or alternatively, data obtained from a public growth graph database may be used to determine recommendations. Further, the type of user-provided information utilized by the system may depend on the type of absorbent article being recommended. By way of example, for feminine hygiene product recommendations, the user-provided information may include, but is not limited to, menstrual frequency, current product used, flow rate, date of previous menstrual cycles, typical cycle length, absorbency of product used, and the like. Additionally or alternatively, growth data and charts may be used as the age and/or weight of the wearer varies significantly. For incontinence products, the information provided by the user may include the type of incontinence, the current product used, etc.

The systems and methods described herein can be used to generate recommendations for a wide variety of absorbent articles, including products in feminine hygiene or adult incontinence spaces, for example. Thus, while many of the figures and examples described herein include infants for illustration purposes, the disclosure is not so limited. For example, for products in feminine hygiene or adult incontinence spaces, the user of the system may be the subject, and an image of the subject may be captured by using a camera timer feature or a mirror. Images processed according to the present disclosure may include a subject, such as a whole-body image, or the subject's undergarment laid flat on a flat surface. As described in more detail below, if the subject is standing and their entire body is present in the image and the height of the person is provided to the system, the scale marker may be the person itself. Alternatively, any other type of scale marker may be included in the image, such as a hand held scale marker that lies flat against the body, a sticker on the body or clothing, or other suitable object of known dimensions. In some embodiments where the subject takes a reflectance image in a mirror, the scale marker may be the phone itself, as it appears in the reflectance.

Referring now to FIG. 1, a user 122 interacting with an exemplary disposable article recommendation computing system 100 according to a non-limiting embodiment is depicted. In fig. 1, a user 122 is seeking an absorbent article recommendation for a subject 124. User 122 may position subject 124 within a field of view 130 of image capture device 114. In the illustrated embodiment, a scale object 128 is also included in the field of view 130 and is captured in the image 118. The size of the scaled object 128 or the size of the printed indicia on the scaled object 128 may be known to the disposable article recommendation computing system 100. In this regard, the scale object 128 may be, for example, standard sized paper slips, credit cards, playing cards, paper currency, coins, and the like. In some embodiments, the package of disposable articles may be sold with the scale object 128. Such a proportional object 128 may be printed or otherwise disposed on the packaging, or may be a separate object contained within the packaging. In some embodiments, the scale object 128 may be downloaded from a website and printed by the user 122. However, as described in more detail below, in other implementations, the scale object 128 need not be included in the image 118.

Once the subject 124 is properly positioned within the field of view 130, an image 118 including representations of the subject 126 and the scale object 128 may be collected. In some embodiments, assistance or guidance may be provided to user 122 to assist in proper alignment and placement of subject 124 within field of view 130 of image capture device 114. The image 118 may be a single image, multiple still images, or a movie or video clip of the subject 124.

To ensure that the image 118 is available for the image processing techniques described herein, the disposable article recommendation computing system 100 may execute various routines upon receiving the image 118. For example, the disposable article recommendation computing system 100 may perform perspective correction to account for any keystone distortion that may be present in the image 118. The amount of correction needed to account for keystone distortion may also be determined and if the amount of correction exceeds some preset boundary, the image may be rejected.

The disposable article recommendation computing system 100 may also execute various error checking routines to ensure that the subject 124 is properly oriented and positioned within the field of view 130. Real-time feedback may be provided to the user regarding suggested adjustments, such as position adjustments or environmental adjustments, which may need to be made before processing may proceed to the next step. For example, subject 124 may need to be supine, with image capture device 114 positioned substantially overhead, and an error checking routine may confirm the orientation. Additionally, the scale object 128 may need to be placed such that all four corners or other attributes of the scale object 128 are visible and properly positioned within the field of view 130. Various error checking routines may also check whether the subject's eyes are open and looking at the image capture device 114, as well as other posture or position aspects of the subject (i.e., a limb that is sufficiently visible and appropriately oriented for analysis). The error checking routine may confirm that the subject 128 is properly positioned within the box such that the subject 128 is a sufficient distance from the edges of the box and is properly centered. Various routines may perform, for example, face tracking, gesture tracking, object tracking, and the like. Some initial routines or subroutines may be performed to increase the overall processing speed of image analysis and subsequent recommendations. For example, by first running a routine that detects the face of the subject 128, the system can then quickly estimate where the subject's body should be located within the frame. The disposable recommendation computing system 100 may also confirm that the illumination level is suitable for image processing, confirm that the subject is in focus, confirm the presence of a scale object, and the like. In the event that the image 118 does not satisfy one or more of these checks or other types of pre-processing checks, the user 122 may be requested to provide additional images 118 in order to correct the problem. For example, some error checking may be performed in real-time, either locally or remotely according to the system configuration, on a real-time preview, when the subject 124 is positioned in the field of view 130. In some implementations, user 122 will not allow images to be collected with image capture device 114 until certain criteria are met. For example, using the live preview, the user 122 may reposition the subject 128 and/or the scale object 128, for example, to satisfy an error checking routine or verification process.

In some embodiments, certain error checking routines or verification processes may be performed locally at the image capture device 114 (such as, for example, by a remote computing device 238 associated with the image capture device 114 shown in fig. 2), and other processes may be performed by the disposable article recommendation computing system 100. For example, one or more routines that may be executed relatively quickly may be executed by a remote computing device. Such high-speed routines may be particularly helpful, for example, in connection with real-time preview screens on remote computing devices, as described in more detail below. Other routines that may require a higher level of accuracy and may require more time and resources may be executed by the disposable article recommendation computing system 100. Further, to increase processing speed or to match the resolution of the images used to train the various machine learning models, the images 118 may first be downsampled for initial processing. For example, during initial processing, corners (or other attributes) of the scaled object 128 may be detected using a low resolution image. It should be appreciated that the various processes associated with the system may best run at different resolutions, as is the case with various machine learning algorithms. Furthermore, some processes may run too slow on full resolution images. In some embodiments, after each algorithm runs at its optimal resolution, the data may be remapped to the original image resolution (or a single downsampled resolution, for example). Various combinations of upsampling, downsampling, cropping, or using portions of an image may be used. Downsampling may also reduce file size, which may advantageously allow downsampled images to be sent between various devices faster than full resolution images. Once the appropriate image 118 is collected and the perspective is corrected, the disposable recommendation computing system 100 may determine the scale of the image 118. While the scale of the image 118 may be determined using a variety of suitable techniques, the scale of the image 118 in FIG. 1 may be determined based on the known dimensions of the scale object 128. For example, knowing that the scale object 128 is an 8.5 inch by 11 inch piece of paper, the disposable recommendation computing system 100 may determine the overall scale of the image 118. Accordingly, a pixel to inch conversion or other suitable scale of the image 118 may be determined by the disposable article recommendation computing system 100.

The disposable article recommendation computing system 100 may perform various processes to identify physical attributes of the subject 124, as represented in the image 118. In some embodiments, for example, joint locations, such as ankle, hip, shoulder, etc., are identified. Using the scale of the image 118, the disposable recommendation computing system 100 may then determine various dimensions of the physical attributes of the subject 124. As provided herein, the various dimensions of the physical property may be determined using any of a variety of techniques, such as linear correlation models, machine learning algorithms, and the like. Non-limiting examples of dimensions that may be determined include, but are not limited to, hip width, waist width, torso length, distance between ears, distance between pupils, and the like.

Once the physical attributes of subject 124 are determined, a plurality of fit parameters may be generated. The fit parameters may be related to one or more parameters used to generate sizing models for various absorbent articles. In fig. 1, the fit parameters generated by the disposable article recommendation computing system 100 include an estimated waist circumference of the subject 124, an estimated thigh circumference of the subject 124, and an estimated rise measurement of the subject 124. However, it should be understood that any of a number of different types of fit parameters may be utilized.

A plurality of sizing models 108 for the prefabricated absorbent articles may be stored within the database 106. In general, each prefabricated absorbent article may have an associated sizing model 108 that includes a range of fit parameters for that particular article. In the illustrated embodiment, for example, each sizing model 108 is a three-dimensional model that includes a waist circumference range, a thigh circumference range, and a rise measurement range associated with a particular absorbent article. In other implementations, the sizing model may utilize different fit parameters. In determining the fit parameters for the subject 124, the disposable article recommendation computing system 100 can apply the fit parameters to the sizing model 108 to determine which pre-made disposable article is appropriately sized for the subject 124. In some cases, the fit parameters of subject 124 may fall within the boundaries of a plurality of different sizing models 108 that are each associated with different sizes of absorbent articles. In such cases, the disposable article recommendation computing system 100 may make recommendations based on which absorbent article is likely to fit better. However, in some configurations, fit parameters or other physical attributes of the wearer may be used to manufacture a customized absorbent article in an on-demand manufacturing process. For example, a customized absorbent article may have a certain base design while allowing certain aspects to be modified and custom configured to fit an intended subject based on the subject's fit parameters. After the customized absorbent article is manufactured, the absorbent article may be shipped directly to the subject's home or to, for example, a retail location adjacent to the subject.

Still referring to FIG. 1, recommendations 120 may be provided to a user 122 via a suitable user interface 116. User interface 116 may be any suitable device or method capable of communicating information to user 122, such as a display screen of a computing device, a text message, an email message, an in-application message, and so forth. The scope of the recommendation 120 may vary. In some implementations, the recommendation 120 identifies a size of absorbent articles suitable for the subject 124, shown as a product size recommendation 134. In some implementations, the recommendations 120 may include additional information, such as product type recommendations 132 and product lineup recommendations 136. This additional information may provide a recommendation as to whether subject 124 should wear, for example, a taped diaper or a training pant diaper. Further, the recommendation 120 may indicate a recommended number of products of a particular size to purchase (i.e., based on expected growth of the subject 124). The recommendation 120 may also indicate when the recommended product is to be sized up (size-up). Depending on the type of product, the recommendation 120 may indicate, for example, the size of the disposable liner or disposable insert. The recommendations 120 may also provide purchasing information (such as identifying an online retailer or physical retailer that sells the recommended products) and/or provide information about subscribed purchasing programs. Generally, a subscription purchasing program may routinely send batches of recommended sized disposable articles to the user 122 over time. In some cases, the size of the disposable articles provided in batches may automatically increase over time to account for the growth of subject 124.

The disposable article recommendation computing system 100 may be provided using any suitable processor-based device or system, such as a personal computer, a mobile communication device, a laptop computer, a tablet computer, a server, a mainframe, or a collection of multiple computers (e.g., a network). The disposable article recommendation computing system 100 may include one or more processors 104 and one or more computer memory units 102. For convenience, only one processor 104 and only one memory unit 102 are shown in FIG. 1. The processor 104 may execute software instructions stored on the memory unit 102. The processor 104 may be implemented as an Integrated Circuit (IC) having one or more cores. The disposable article recommendation computing system 100 may also utilize one or more Graphics Processing Units (GPUs) to assist various aspects of image processing to execute one or more machine learning models and/or convolutional neural network models suitable for visual imaging analysis. The memory cells 102 may include volatile and/or nonvolatile memory cells. Volatile memory units may include, for example, Random Access Memory (RAM). The non-volatile memory unit may include, for example, Read Only Memory (ROM), as well as mechanical non-volatile memory systems such as, for example, hard disk drives, optical disk drives, and the like. The RAM and/or ROM memory units may be implemented as, for example, discrete memory ICs.

The memory unit 102 may store executable software and data for use by the disposable article recommendation computing system 100 described herein. When the processor 104 of the disposable article recommendation computing system 100 executes the software, the processor 104 may be caused to perform various operations of the disposable article recommendation computing system 100, such as analyzing the image, determining physical attributes and fit parameters, comparing the fit parameters to a sizing model, and providing recommendations to the user.

The data used by the disposable article recommendation computing system 100 may come from various sources, such as a database 106, which may be, for example, an electronic computer database. The data stored in database 106 may be stored in non-volatile computer memory, such as a hard disk drive, read-only memory (e.g., ROM IC), or other types of non-volatile memory. In some embodiments, one or more databases 106 may be stored, for example, on remote electronic computer systems. It should be understood that a variety of other databases or other types of memory storage structures may be utilized or otherwise associated with the disposable article recommendation computing system 100.

According to various embodiments, the disposable article recommendation computing system 100 may include one or more computer servers, which may include one or more web servers, one or more application servers, and/or one or more other types of servers. For convenience, only one web server 110 and one application server 112 are depicted in FIG. 1, but those of ordinary skill in the art will appreciate that the present disclosure is not so limited. The

servers

110, 112 may be comprised of a processor (e.g., CPU), memory units (e.g., RAM, ROM), a non-volatile storage system (e.g., hard drive system), and other elements.

In some embodiments, the web server 110 may provide a graphical web user interface, such as the user interface 116, through which various users 122 may interact with the disposable article recommendation computing system 100. The graphical web user interface may also be referred to as a client portal, client interface, graphical client interface, and the like. The web server 110 may accept requests from various entities (such as HTTP/HTTPs responses), such as HTTP/HTTPs requests, as well as optional data content, such as web pages (e.g., HTML documents) and link objects (such as images, videos, etc.). The application server 112 may provide a user interface, such as interface 116, for a user that does not use a web browser to communicate with the disposable article recommendation computing system 100. Such users may have special software installed on the user's computing device to allow the user to communicate with the application server 112 via a communication network, as described in more detail below. Further, the user interface may include a connection to an automated agent or human agent through any suitable communication portal, such as chat, voice, or video, to guide the user in using the recommendation computing system.

The user 122 may be presented with an interface 116, as generated by the disposable article recommendation computing system 100. User 122 may utilize, for example, a mobile phone, smart phone, tablet computer, laptop computer, desktop computer, kiosk, or other computing device capable of displaying interface 116. As provided above, the interface 116 may identify one or more recommendations 120 based on the image analysis and processing described above.

An alternative embodiment of a disposable article recommendation computing system 200 is illustrated in FIG. 2 and may be similar or identical in many respects to the disposable article recommendation computing system 100 shown in FIG. 1. For example, the disposable article recommendation computing system 200 may include a memory unit 202, a processor 204, and a database 206 for storing a sizing model 208. The disposable article recommendation computing system 100 may include various software programs such as system programs and applications to provide computing capabilities in accordance with the described embodiments. The system programs may include, but are not limited to, an Operating System (OS), device drivers, programming tools, utilities, software libraries, Application Programming Interfaces (APIs), and the like. Exemplary operating systems may include native operating systems and utilize cloud-based computing services, such as MICROSOFT AZURE servers, AMAZON WEB Services (AWS), alibas clouds, and so forth.

The disposable article recommendation computing system 200 may also include various servers, such as a web server 210 and/or an application server 212. As shown in FIG. 2, a user 222 may interact with the disposable article recommendation computing system 200 via a remote computing device 238 having a camera 214 and a user interface 216. The remote computing device 238 may be any type of computer device suitable for communicating over a network, such as a wearable computing device, a mobile phone, a tablet computer, a device that is a combination of a handheld computer and a mobile phone (sometimes referred to as a "smart phone"), a personal computer (such as a laptop computer, a netbook computer, a desktop computer, etc.), or any other suitable networked communication device, such as, for example, a Personal Digital Assistant (PDA), a mobile gaming device, or a media player.

In some embodiments, the remote computing device 238 may provide a variety of applications to allow the user 222 to use the disposable article recommendation computing system 200 to accomplish one or more specific tasks. Applications may include, but are not limited to, web browser applications (e.g., INTERNET EXPLORER, MOZILLA, FIREFOX, SAFARI, OPERA, NETSCAPE NAVIGATOR), telephony applications (e.g., cellular, VoIP, PTT), networking applications, messaging applications (e.g., email, IM, SMS, MMS, blackberry messenger, WeChat, Small Red book, WhatsApp), and so forth. The remote computing device 238 may include various software programs such as system programs and applications to provide computing capabilities in accordance with the described embodiments. The system programs may include, but are not limited to, an Operating System (OS), device drivers, programming tools, utilities, software libraries, Application Programming Interfaces (APIs), and the like. Exemplary operating systems may include, for example, but are not limited to, PALM OS, MICROSOFT OS, APPLE OS, ANDROID OS, UNIX OS, LINUX OS, SYMBIAN OS, EMBEDIDIX OS, Binary Runtime Environments for Wireless (BREW) OS, JavaOS, Wireless Application Protocol (WAP) OS, and cloud-based computing services such as MICROSOFT AZURE Server, AMAZON WEB Services (AWS), ALIBABA cloud, and the like.

The remote computing device 238 may include various components for interacting with the disposable article recommendation computing system 200. Remote computing device 238 may include components for use with one or more applications, such as a stylus, a touch-sensitive screen, keys (e.g., input keys, preset and programmable hot keys), buttons (e.g., action buttons, multidirectional navigation buttons, preset and programmable shortcut buttons), switches, a microphone, a speaker, an audio headset, a depth sensor, an IR projector, a stereo camera, a gyroscope, an accelerometer, and so forth. The user 222 may interact with the disposable article recommendation computing system 200 via a variety of other electronic communication techniques, such as, but not limited to, HTTP requests, in-application messaging, and Short Message Service (SMS) messages. The electronic communication may be generated by a dedicated application executing on the remote computing device 238 or may be generated using one or more applications that are substantially standard to the remote computing device 238. The application may be comprised of or implemented as executable computer program instructions stored on a computer readable storage medium, such as volatile or non-volatile memory, that are retrievable and executable by a processor to provide operations for remote computing device 238. The memory may also store various databases and/or other types of data structures (e.g., arrays, files, tables, records) for storing data for use by the processor and/or other elements of the remote computing device 238.

Similar to the process depicted in fig. 1, the user 222 may collect the image 218 of the subject 224 by placing the subject 224 in the field of view 230 of the camera 214. For example, in the context of a mobile phone, the camera 214 may be a rear-facing camera that provides a preview of an image on the user interface 216. As shown, a scale object 228 may be included in the image 218 for image processing by the disposable article recommendation computing system 200. In the illustrated embodiment, the image 218 including the representation of the subject 226 is provided to the disposable article recommendation computing system 200 via the electronic communication network 240. The communication network may include a plurality of computer and/or data networks (including the internet, LAN, WAN, GPRS network, LTE network, etc.) and may include wired and/or wireless communication links. In some embodiments, the remote computing device 238 may perform various types of pre-processing, such as error checking routines, before providing the image 218 to the disposable article recommendation computing system 200. Further, as described in more detail below with reference to fig. 3-6, in some embodiments, the remote computing device 238 may perform image processing on the image 218 and provide an output of the image processing to the disposable recommendation computing system 200.

Still referring to FIG. 2, upon receiving the image 218, the disposable article recommendation computing system 200 may perform the processing described in FIG. 1 to generate a recommendation 220 for the user 222. The recommendation 220 may then be sent via the network 240 for display on the user interface 216 of the remote computing device 238. The scope, format, and content of the recommendations 220 may vary. In the illustrated embodiment, the recommendation 220 includes a recommended product type 232, a recommended product size 234, and a recommended product lineup 236, although the disclosure is not so limited.

Referring now to fig. 3A-3B, side views of an exemplary scale object 328 and subject 326, each positioned adjacent to an image capture device 314 having a field of view 330, are shown. In both fig. 3A-3B, the scale object 328 is shown placed on the same plane as the plane on which the subject 326 is supine, which is referred to herein as the scale object focal plane 344. The proportional object focal plane 344 may be substantially coplanar with, for example, a floor, a bed, a crib, or other surface on which the subject 326 is placed. Because of the volume of subject 326, the hip of subject 326 is not coplanar with the proportional object focal plane 344. In contrast, the hip of subject 326 is located in a hip focal plane 342 that is substantially parallel to the proportional object focal plane 344, but is closer to the image capture device 314 than the proportional object focal plane 344.

The scale of objects coplanar with the scale object focal plane 344 is determined based on a determination of the scale of the image collected by the image capture device 314 by the scale object 328. Thus, depending on the distance between subject 326 and image capture device 314, this ratio may not provide sufficient accuracy to determine the ratio of hip focal plane 342. However, according to various embodiments of the present disclosure, the distance between proportional object focal plane 344 and hip focal plane 342 may be calculated, estimated, or measured and then used to determine the proportion of hip focal plane 342, thereby allowing a more accurate determination of the size of subject 326 by image analysis. It should be noted, however, that according to other embodiments, the sizing recommendation process described herein may be based on the scale of the proportional object focal plane 344 without regard to any possible proportional difference between the proportional object focal plane 344 and the focal plane in which the physical properties of the subject may lie.

As image capture device 314 is pulled further away from subject 326, the difference between the proportions of hip focal plane 342 and proportional object focal plane 344 will decrease. In FIG. 3A, the image capture device 314 is relatively close to the subject 336, where the distance between the image capture device 314 and the scaled object focal plane 344 is shown as distance D_1AAnd the distance between image capture device 314 and hip focal plane 342 is shown as distance D_2A. The distance between hip focal plane 342 and proportional object focal plane 344 is shown as distance D_3AThe distance D_3AIs equal to D_1ADecrease D_2A. In FIG. 3B, the image capture device 314 is positioned farther away, and the distance between the image capture device 314 and the scaled object focal plane 344 is shown as distance D_1BAnd the distance between image capture device 314 and hip focal plane 342 is shown as distance D_2B. The distance between hip focal plane 342 and proportional object focal plane 344 is shown as distance D_3BThe distance D_3BIs equal to D_1BDecrease D_2B. As the image capture device 314 is pulled away from the subject 326 (i.e., moved from the position shown in FIG. 3A to the position shown in FIG. 3B), a distance D₃Becomes a distance D₁And the difference in the proportions of hip focal plane 342 and proportional object focal plane 344 decreases. Thus, if the image capture device 314 is positioned at a sufficiently far distance from the subject 326, the difference in the scale between the scale object 328 and the subject 326 does not substantially affect the size recommendation and may be considered negligible. However, in the event that the difference in scale is not negligible, the image processing described herein may take into account the distance between hip focal plane 342 and scale object focal plane 344 and compensate for the difference, thereby avoiding situations where the calculated size of subject 326 is determined to be larger than the actual physical size.

A variety of suitable techniques may be used to compensate for the difference between the proportions of hip focal plane 342 (as determined by scaling object 328) and scaling object focal plane 344. It should be noted, however, that some embodiments of the present disclosure do not attempt to compensate for any difference between the proportions of hip focal plane 342 and proportional object focal plane 344, regardless of the magnitude of the difference. If it is desired to compensate for the proportional difference, the proportional difference between hip focal plane 342 and proportional object focal plane 344 is linearly proportional to the distance between hip focal plane 342 and proportional object focal plane 344 to node 332 (figure 3C) of image capture device 314. As understood in the art, a node is a point where all rays entering the lens of the image capture device 324 converge. Further, the focal length provides the distance between the node and the image sensor of the image capture device 314. The focal length of the image capture device 314 may be provided in metadata in the captured image. Using the focal length of image capture device 314, the projected size of subject 326 (as if the projected size of the subject were coplanar with proportional object focal plane 344), and the distance between proportional object focal plane 344 and hip focal plane 342, the proportion of hip focal plane 342 may be determined.

Referring now to fig. 3C, a cross-sectional end view of the torso of subject 326 is shown, with hip 348 schematically illustrated. Hip 348 is located in hip focal plane 342. Distance D_1CMay be determined based on the pixel size of the image on the image capture device 314 and the focal length of the image capture device 314 (as provided in the image metadata) because the size of the scale object 328 is known (i.e., 81/2 inches by 11 inches, or other suitable size). In some embodiments, if distance D_1CConsidered to exceed a certain threshold, the size of subject 326 may be determined without any adjustment to the separation of proportional object focal plane 344 and hip focal plane 342. However, in some embodiments, further processing may be performed to provide such adjustments in order to more accurately determine various dimensions of subject 326. In one embodiment, the actual hip width (i.e., dimension H in figure 3C)_A) The hip width H that may be based on the subject's projection on the proportional object focal plane 344 size as collected by the image capture device 324_PPixel size and distance D_1CTo be derived. For example, the hip width H may be based on the projection onto the proportional object focal plane 344 of the subject 326_PThe distance of hip focal plane 342 above proportional object focal plane 344 (shown as distance D) is determined using the hip plane constants_3C). In an exemplary implementation, the thickness T of the subject 326 is considered to be the hip dimension H of the subject_AAbout 90% of the total. Further, in a side view, the hip of the subject is considered to be located in the middle of the height of the subject (i.e., 50% of thickness T). Thus, the subject's hip may lie in a plane that is about 0.45 times the hip width above the proportional object focal plane 344, with the hip plane constant of 0.45 being calculated as the product of 0.9 and 0.5. Thus, the distance D_3CCan be estimated as hip size H_A0.45 times of. Once the distance D is determined_3CThe proportions of hip focal plane 342 may be extrapolated linearly according to the proportions of proportional object focal plane 344, as determined based on proportional object 328. It should be understood that while a hip plano constant of 0.45 is provided herein, other hip plano constants may be used without departing from the scope of the present disclosure.

Furthermore, other methods may be used to measure the distance between hip focal plane 342 and proportional object focal plane 344, according to various implementations. For example, the distances to hip focal plane 342 and proportional object focal plane 344 may be measured directly by an instrument associated with image capture device 314. Such distances may be measured or at least interpolated by stereophotogrammetry, infrared triangulation, laser ranging, infrared time of flight, or combinations thereof. In some implementations, such as when the remote computing device has multiple image capture devices 314, multiple images taken simultaneously at multiple locations may be analyzed. In any case, any of a variety of methods may be used in accordance with the systems and methods described herein to accurately account for vertical differences between the plane in which the scale object 328 is placed and the plane of the physical feature of the subject 326. Fig. 4-7 schematically depict various exemplary operating arrangements of a disposable article recommendation computing system according to the present disclosure. Referring first to fig. 4, an operational arrangement similar to that depicted in fig. 2 is shown. That is, the remote computing device 438 collects the image 418 using the camera 414. The image 418 is provided to the disposable article recommendation computing system 400 via an electronic communication network 440. The disposable article recommendation computing system 400 is configured to perform image processing 442 to determine fit parameters of the subject in the image 418. The disposable article recommendation computing system 400 is further configured to execute the product fit process 444 to apply the fit parameters of the subject in the image 418 to the sizing model 408. After the image processing 442 and the product fit processing 444, the disposable article recommendation computing system 400 may transmit the recommendation 420 for display on the user interface 416 of the remote computing device 438.

Fig. 5 depicts an embodiment having an alternative operational arrangement. As shown, the remote computing device 538 includes a camera 514 and a user interface 516, similar to the previously described embodiments. However, in this embodiment, the sizing model 508 is stored by the remote computing device 538. Additionally, image processing 542 and product fit processing 544 are performed by remote computing device 538. Thus, for example, the remote computing device 538 may generally provide the functionality of the disposable article recommendation computing system 100. As shown, in some embodiments, the remote computing device 538 may provide various data to the disposable article recommendation computing system 500 over the network 540. In the illustrated embodiment, performance data 546 is provided to disposable article recommendation computing system 500, and various updates 548 can be provided to remote computing device 538. Performance data 546 may be collected from a user of remote computing device 538 and may relate to a recommended quality of the absorbent article based on, for example, product fit or product performance. Performance data 546 as collected from multiple users may be used to change and update various fit parameters over time, such as using a self-learning model. Such updated fit parameters may then be provided to the remote computing device 538 in an attempt to continually improve the product fit recommendations over time. Fig. 6 depicts yet another exemplary alternative operational arrangement. In this embodiment, the remote computing device 638 has a camera 614 and a user interface 616. The remote computing device 638 is configured to perform image processing 642 on images collected by the camera 614. As a result of this image processing 642, fit parameters 650 may be determined and provided to the disposable article recommendation computing system 600 via communication over the communication network 640. Upon receiving fit parameters 650, disposable article recommendation computing system 600 may execute product fit process 644 to apply the received fit parameters 650 to sizing model 608. The disposable article recommendation computing system 600 may then provide the recommendation 620 to the remote computing device 638 for display on the user interface 616. Thus, fig. 6 depicts an exemplary operational arrangement in which process steps are split between the remote computing device 638 and the disposable-article recommendation computing system 600.

FIG. 7 depicts another embodiment having an alternative operational arrangement that splits the process between a remote computing device and a disposable article recommendation computing system. In this exemplary arrangement, the sizing model 708 is stored by a remote computing device 738. The camera 714 of the remote computing device is used to collect an image 718 that is provided to the disposable article recommendation computing system 700 over the communication network 740. Disposable article recommendation computing system 700 is configured to perform imaging process 742 in order to determine fit parameters 750. The fit parameters 750 may then be communicated to the remote computing device 738 for the product fit process 744. Product fit process 744 may apply fit parameters 750 to sizing model 708 and generate recommendations to display on user interface 716.

As shown in fig. 3-6, some of the processing may occur locally on the remote computing device, while other processing may occur at the disposable recommendation computing system. Thus, relatively fast algorithms executable on the remote computing device may be separated from relatively slow algorithms executable on the server side. The method may allow real-time preview guidance of the user in real-time, for example, because such routines that support real-time preview guidance functionality may be executed locally by the remote computing device. Thus, with respect to identifying the scale object and determining the scale of the image, a combination of processes may be used that includes low resolution object detection algorithms that first determine the region in the image where the scale object is located. For example, such processes may be performed locally at a remote computing device. Such processes may also include high resolution object detection algorithms that attempt to identify the best search neighborhood for a particular feature of the scaled object, such as a corner, edge, printed mark, etc. The high resolution algorithm or process may be run on a disposable article recommendation computing system. These processes may also include various refinement algorithms, such as edge detection algorithms, to accurately locate features of the scaled object. These algorithms may be combined in various ways including feeding partial images from one algorithm to the next (i.e., a directed search), weighting the search parameters of one algorithm against the other based on the input to the other algorithm, or selecting the results of a single algorithm based on its confidence in locating a feature of interest.

Thus, in some implementations, a set of ultra-low resolution algorithms may be run on a remote computing device. More particularly, these algorithms may be fast enough to execute on a remote computing device while providing sufficient boot-up and/or error checking. However, these algorithms may not necessarily be accurate enough to determine the body size of the subject. The goal of these algorithms is to provide additional error checking while also providing guidance for high resolution algorithms. For example, a high resolution algorithm executed by the disposable article recommendation computing system may accurately select points of interest in an image where the search is guided by low resolution results. In some cases, the high resolution algorithm is an algorithm that may be accurate enough to be used as a basis for a model.

For illustrative purposes only, the following provides exemplary interactions of various processes according to one non-limiting embodiment. First, appropriate programming functions may be performed by the remote computing device to locate the scale object. In some embodiments, an open source computing vision library such as OpenCV may be used for this process. This process may be used to guide the user to place the scale object in the appropriate position in the box, to error check the presence of the scale object, etc. The process may use simple object detection, e.g., placing a bounding box around the scaled object, although the process is not necessarily designed to find edges or corners or perform image segmentation. Next, at the disposable article recommendation computing system, a first pass with a machine learning algorithm may confirm the presence of uncreped white paper or other suitable scale object (i.e., error checking) and locate a search neighborhood for corners or other attributes, as the case may be. An edge detection algorithm may then be executed by the disposable article recommendation computing system within the corner search neighborhood to accurately select the scaled object corners or other attributes. It should be noted that various edges appear in the whole image, which is why it is beneficial for the algorithm to focus on a specific area where the scale object appears. Once identified, these locations can be used to scale the image according to the present disclosure.

In any case, by sharing various image processing algorithms between the local remote computing device and the disposable article recommendation computing system, the overall accuracy and speed of the size recommendation process may be optimized while also allowing scalability to be achieved in a large number of simultaneous users. Referring now to FIG. 8, an exemplary process that may be performed by the disposable article recommendation computing system 800 is schematically illustrated. The disposable article recommendation computing system 800 may receive an unprocessed image 818A that includes representations of the subject 826 and the scale object 828. During image processing 842, various error checks may be performed on unprocessed images 818A to ensure that the images are suitable for analysis. Image processing 842 may also correct for any keystone effect, which is distortion caused by the relative angle between the image capture device and the subject. Any suitable method for keystone correction may be utilized including utilizing information collected by sensors associated with the camera (such as accelerometers, gyroscopes, etc.) to assist in correcting the angle of the camera. Image processing 842 may make any other image corrections that may be needed, such as, for example, corrections or adjustments to light, color, intrinsic camera parameters, lens distortion, or other distortions. Further, the amount of correction required for the image may be quantified for the image such that if correction is required that exceeds a threshold that would affect the recommended accuracy, a new image is requested.

With specific regard to performing the image processing routine on the scale object according to various embodiments, various error checking routines may be performed to ensure that the scale object included in the image is sufficient and usable by the system. For example, confidence levels for various aspects of scaled object recognition may be used. In some embodiments, confidence levels for corner detection of a sheet or sheet edge are ascertained to determine whether the scale object in the image is acceptable. Additionally or alternatively, various algorithms may be used, for example, to determine whether an appropriate scale object is present and to search for other factors in addition to the edges or corners of the scale object. In one implementation, image analysis is performed to determine whether the scale object includes print or lines in an attempt to determine whether the scale object is actually suitable for processing. The image analysis may also detect the presence of holes (such as 3 ring adhesive holes or spiral adhesive holes) in the scale object. The image analysis may also check, for example, whether corners may not be visible due to tearing, or look for evidence of folding or bending of a scale object. It may be determined whether the scale object is not available and another image with the appropriate scale object is requested, or in some cases, the image is not allowed to be collected until the problem with the scale object is resolved. Additionally, algorithms according to the present disclosure may search for a linear or otherwise conforming edge of the scale object to a desired shape (e.g., rectangular or circular) to determine whether the scale object has been folded or otherwise deformed. The algorithm may also search the texture to detect wrinkles in the surface of the scaled object. Based on the results of the algorithm associated with the scale object, the system may determine whether to continue processing the remaining images or reject the images and request a different image from the user. In some embodiments, the user may be provided notification of an inference about rejection (i.e., missing corners, detected wrinkles, incorrect aspect ratio, etc.). In addition to error checking and verification routines associated with the scale object, a variety of similar routines may be performed with reference to analysis of the subject. For example, in some embodiments, the confidence value for each joint position and/or the number of joint positions predicted for a single joint may be used to detect errors and determine whether an image is suitable. Thus, according to the systems and methods described herein, a variety of error checks may be performed at various points during processing to ensure that the image is suitable.

After any keystone distortion, in addition to any other problems, is corrected and any other error checks are satisfied, the processed image 818B may be analyzed to determine the scale of the image. In some implementations, the detected scale attribute 852 of the scale object 828 is utilized to determine the scale of the processed image 818B. For example, if the proportional object 828 is a conventional piece of paper that measures 8.5 inches by 11 inches, the disposable-article recommendation computing system 800 may identify the proportional object 828 and then measure the corner-to-corner dimensions of the proportional object 828 during the imaging process 842. The disposable article recommendation computing system 800 may also confirm that the two corner-to-corner dimensions are similar, which confirms that the correction for any keystone effect is successful. The disposable recommendation computing system 800 may then determine the scale of the processed image 818B based on the known corner-to-corner dimension of the paper sheet of 13.9 inches. In some embodiments, analyzing the aspect ratio of the scale object 828 before and/or after correction, or similarly, analyzing any dimension of the scale object (e.g., first side length, second side length, width, diagonal, etc.) before and/or after correction, may be used to detect images that are not suitable for processing recommendations. In some embodiments, unprocessed image 818A may include more than one scale object 828, such that one of the scale objects in the image is used for perspective correction and the other scale objects in processed image 818B may be used to confirm that the correct scale is applied and that the perspective is sufficiently removed. Further, while FIG. 8 depicts the use of a conventional paper sheet for the scale object 828, the invention is not so limited. In addition, other measurements of the scale object 828 may be used to determine scaling. For example, in some implementations, the scale attributes 852 may be the measured top and bottom widths of the scale object 828 and the left and right heights of the scale object 828. The top and bottom widths of the scale object 828 may be averaged to determine the scale in the X direction, and the left and right heights of the scale object 828 may be averaged to determine the scale in the Y direction. Determining both the X-direction scale and the Y-direction scale of processed image 818B may increase overall accuracy, particularly when the pixels of processed image 818B are not square, such as in the case of a video image.

Processed image 818B also schematically shows an exemplary physical attribute 854 of subject 826 that may be ascertained during image processing 842. In the illustrated embodiment, various joint positions and physical features (i.e., ear position, eye position, nose position, etc.) are determined. Such a determination may be performed by the disposable article recommendation computing system 800, or the determination may be performed locally, remotely by a third party computing system, or by any other suitable resource. In some implementations, a machine learning based model is utilized to identify joint positions or other types of physical features. An exemplary resource for determining the joint location of a subject image is provided by openpos, University of camerameron, Pittsburgh, Pennsylvania, of Pittsburgh, Pennsylvania. Openpos is a real-time multi-person system for collectively detecting human, hand, face, and foot keypoints on a single image. Other techniques for detecting physical attributes 854 of a user may include edge detection and object detection algorithms. By way of example, as an alternative to or in addition to determining joint positions, an algorithm may be used to identify the triangle formed by the nipple and navel of subject 826. Based on the dimensions of this triangle as determined based on the scale of the image, a model may be applied to determine the appropriate sizing of subject 826.

Once the physical attributes 854 have been identified, the disposable article recommendation computing system 800 can determine various sizes of the subject 826 based on the scale. While the determined dimensions may vary based on the physical properties 854, in some implementations, exemplary dimensions include torso measurements, distance between ears, shoulder width, and hip width.

During the product fit process 844, various fit parameters may be determined by the disposable article recommendation computing system 800 based on the dimensions associated with the physical attributes 854. The particular fit parameters utilized by the disposable article recommendation computing system 800 may then be selected to match the fit parameters of the various sizing models 808. For example, sizing models 808 may each be defined by an acceptable range of certain fit parameters, such as, but not limited to, waist circumference at the navel, rise measurements from the navel to the back, and thigh circumference. Accordingly, the disposable article recommendation computing system 800 may estimate fit parameters based on the determined dimensions of the physical attributes 854. In some embodiments, for example, a statistical model is used to correlate various dimensions of the physical attribute 854 to various fit parameters or directly to a preferred diaper size. Further, in some embodiments, in addition to the various dimensions of the physical attributes 854 ascertained by image analysis, other inputs to the statistical model may include gender and age, as well as other specific dimensions or information entered by the user, as described in more detail below.

Referring now to FIG. 9, exemplary fit parameters 956 based on physical attributes 954 are shown. In the illustrated embodiment, fit parameters 956 include waist circumference at the navel, rise measurements from the navel to the back, and thigh circumference. Similar to fig. 8, fit parameters 956 are based on physical attributes 954 of the subject, shown as joint position, eye position, etc., and scale object 928. However, fig. 10-11 illustrate that other methods may be used to determine the scale of the image. Referring to FIG. 10, an exemplary process for determining the scale and fit parameters is based on body part size, as provided by a user of the disposable article recommendation computing system. For example, in the illustrated embodiment, foot size 1054 is provided by the user to the disposable recommendation computing system. Foot size 1054 may be provided in any suitable format, such as shoe size or foot length. When the subject's foot is captured within the image at an appropriate angle, the disposable recommendation computing system may utilize the known dimensions of the subject's foot to determine the scale of the image. Based on the proportion and physical attributes of the subject, various fit parameters 1056 can be determined. FIG. 11 depicts another configuration in which a user provides a head circumference to a disposable article recommendation computing system. Similar to fig. 10, the disposable recommendation computing system may determine the scale of the image using known dimensions of the subject or parameters as calculated from known head circumference, such as head width 1154. In turn, the disposable recommendation computing system may then determine the fit parameters 1156. In other embodiments, one or more of the fit parameters, such as waist size or height, may additionally or alternatively be provided by the user in order to scale the image. For example, with respect to height, the height of the subject in the image may be calculated from the sum of the joint distances. Furthermore, interocular distance or eye diameter can be used to scale the image, as these sizes have very small differences in the population and only change very slowly with age. By way of example, the diameter of the eyeball of a newborn is 16mm +/-2mm, and the diameter of the eyeball at the age of 3 is 19mm +/-2 mm. Thus, upon receiving the age of the subject, the disposable recommendation computing system may utilize the age versus eyeball diameter curve to identify an eyeball diameter within a few millimeters of the correct scale.

Other technique or techniques, such as photogrammetry, may also be used to determine the scale of the image. Two or more images of the subject may be taken from two or more known locations, with the size of the object within the photograph being calculated by tracking the reference point. Alternatively, multiple lenses on a single device may capture images that may be used in a photogrammetric process. Another approach is to utilize a single camera that moves in space and measure the position of the camera movement. Alternatively, a point cloud or structured light may be used. In this method, an array of known patterns or points of the bars may be projected onto the target area/subject. Then, when the projection light falls on a different 3D object, the camera can read the deformation of the projection light. Using algorithms, the 3D shape and size of those objects can be calculated based on the measured deformations of the known light patterns. Structured light may be visible, IR, or other wavelengths as long as a camera configured for this purpose can detect the anamorphic light pattern. Alternatively, a depth sensor may be used to calculate the distance to the object and calculate the scale based on the number of pixels a relatively flat object (such as an infant or a portion of an infant's body) occupies in the image and using triangulation or trigonometric functions to calculate the scale. Alternatively, a tilt sensor, position sensor, accelerometer, or other sensor from a smartphone or other device may be incorporated into the method to help calculate the scale in the acquired image or 3D scan.

Referring now to fig. 12, a simplified sizing model 1208 of sizes 1-6 of an exemplary product lineup for a disposable article is schematically depicted. While each of the sizing models 1208A-1208F is defined in terms of waist size and rise size, the present disclosure is not so limited. For example, the sizing models 1208A-1208F may include additional dimensions (such as thigh girth) or may be based on different dimensions. In any case, in accordance with the present disclosure, the disposable article recommendation computing system may apply the fit parameters of the subject as determined by image analysis to the sizing model 1208 to determine which one or more sizes of disposable articles may be worn by the subject.

Fig. 12 graphically illustrates exemplary fit parameters for two different subjects, shown as subject 1260 and subject 1262. Referring first to the fit parameters defined by subject 1260, a disposable article recommendation computing system can determine that subject 1260 falls within size 1 disposable article sizing model 1208A and provide such recommendation to the user. However, the fit parameters defined by subject 1262 cause subject 1262 to fall within multiple sizing models (i.e., sizing model 1208B for size 2 and sizing model 1208C for size 3). The disposable article recommendation computing system may use various methods to determine which of these two sizes to recommend to the user. According to the illustrated embodiment, the disposable recommendation computing system determines the relative distance from the fit parameters defined by the subject 1262 to each of the boundaries of the sizing model involved. In the illustrated embodiment, distances 1266a-d are distances to the boundaries of size-setting model 1208C of size 3, and distances 1264a-d are distances to the boundaries of size-setting model 1208B of size 2. More particularly, distances 1266a and 1266C are distances to the upper and lower boundaries of the waist fit parameter of size 3 sizing model 1208C. Distances 1266b and 1266d are distances from the upper and lower boundaries of the crotch fit parameter of size setting model 1208C of size 3. Distances 1264a and 1264c are distances to the upper and lower boundaries of the waist fit parameter of size 2 sizing model 1208B. Distances 1264B and 1264d are distances to the upper and lower boundaries of the crotch fit parameter of size setting model 1208B of size 2.

Once the distances 1266a-d and 1264a-d are determined, the disposable article recommendation computing system may recommend a disposable article size based on the measured distances. In some embodiments, for example, the disposable article recommendation computing system may identify a minimum dimension of all boundary dimensions 1266a-d and 1264a-d for each sizing

model

1208B and 1208C. The disposable article recommendation computing system may then compare the two minimum dimensions and recommend a sizing model having the largest of those dimensions.

Fig. 13-20 depict simplified exemplary user interface displays on various computing devices according to non-limiting embodiments. The user interface display may be, for example, a component of the disposable article recommendation computing system 100 of fig. 1 or any of the

remote computing devices

238, 438, 538, 638, 738 of fig. 2 and 4-6. Referring first to FIG. 13, a series of interfaces 1300A-1300B are shown that can be presented on a computing device 1320. Interface 1300A shows an exemplary image preview pane 1302 during the image collection process. The image preview pane 1302 may be used to properly align the subject in the field of view. In some embodiments, image preview pane 1302 may include a graphical guide area or other type of feedback to assist user alignment, such as augmented reality. As shown, the interactive element 1304 may be activated by a user to take a picture or series of pictures of the subject. Images collected by computing device 1320 may be processed in accordance with the present disclosure and interface 1300B may be presented to a user. Interface 1300B may present recommendations 1306 in any suitable format. It should be appreciated that the recommendation 1306 may include a variety of information associated with the recommendation, such as a recommended product type, a recommended product lineup, and so forth. In some embodiments, interface 1300B may also include fit metrics 1308. Fit metric 1308 can indicate a recommended manner in which the disposable article is expected to fit the subject. Fit metric 1308 can be based on, for example, the relative position of a fit parameter of a subject within a sizing model of a recommended size disposable article.

FIG. 14 depicts other exemplary simplified interfaces 1400 that may be presented on a computing device 1420 by a disposable article recommendation computing system. Interface 1400A shows an exemplary image preview pane 1402 during the image collection process. Image preview pane 1402 may be used to properly align the scale object 1428 in the field of view and guide the user to collect images at a sufficient distance from subject 1426. In this exemplary implementation, image preview pane 1402 includes a scaled object graphical guide area 1429 for assisting the user in the image collection process. As shown in interface 1400B, upon detecting proper alignment of the scale object 1428 and the scale object graphical guide area 1429, the interactive element 1404 may be presented to the user. The user is allowed to collect images only after the scale object 1428 is properly positioned within the boundaries of the scale object graphical guide region 1429. Thus, the use of the scaled object graphical guide region 1429 may provide real-time feedback to the user to aid in the image collection process.

The use of the scaled object graphical guide 1429 may ensure that the subject 1426 is a sufficient distance from the computing device 1420. By forcing the user to collect images from a particular distance, the necessity of interpolating the scale of subject 1426 according to fig. 3A-3C may be reduced or eliminated. Additionally, the scale object graphical guide 1429 may also ensure that the relative angle between the computing device 1420 and the scale object 1428 is acceptable for processing, thereby avoiding the need to over-correct image keystone distortion due to perspective effects.

The scale object graphical guide region 1429 may also direct the user to place the scale object 1428 in a particular orientation relative to the subject 1426. Referring to fig. 14, the placement of the proportional object graphical guide area 1429 in the image preview pane 1402 ensures that the proportional object 1428 is placed to the upper left of the subject 1426. Guiding the user to place the scale object 1428 in a particular location may reduce the overall variability of the collected image. The reduced variability may improve overall accuracy because the relative positioning of the scale object 1428 and the subject 1426 in the image may be selected to match the relative positioning of the images used to train the system and develop the model. Additionally or alternatively, a graphical guide region may be used to assist in alignment and placement of the subject. Such graphical guide regions may direct the user to place the subject in a particular orientation (i.e., head towards the top of the frame and legs towards the bottom of the frame), which may help reduce variability in, for example, the collected images.

FIG. 15 depicts another exemplary simplified interface 1500 that may be presented by a disposable article recommendation computing system on a computing device 1520. Interface 1500 may provide a disposable consumption prediction 1510 as determined by a disposable recommendation computing system. The disposable article consumption prediction 1510 may include a variety of information, such as a predicted consumption rate (i.e., the number of absorbent articles used over a period of time) and/or the amount of time until the wearer will need to move to a different size and/or product array. In the illustrated embodiment, for example, disposable consumption prediction 1510 includes a recommended size 1506, a time period prediction 1512, and a quantity prediction 1514 based on predicted consumption rates. Thus, disposable article consumption prediction 1510 informs the user that the intended subject is wearing a size 2 disposable article for the next 22 days, and 65 articles will be consumed over the time period. The model used to determine such disposable product consumption prediction 1510 can be based on any number of inputs, including the physical attributes of the subject, age, gender, growth history, inferred disease control and prevention center (CDC) growth charts, and the like. For example, some inputs may be determined based on image analysis, while other inputs may be based on user-supplied information, such as information provided at the time of image submission or information previously submitted at the time of creation of the user profile. In some embodiments, a series of images of the subject may be collected over time (i.e., greater than 3 days), wherein the growth pattern of the subject is determined based on the growth image-to-image of the subject. Various techniques may be used with respect to determining the predicted consumption rate. For example, a user may enter information regarding the time they purchased absorbent articles and the number of times they purchased. This information may be collected using any suitable technique, such as a user manually providing information through an interface or scanning a UPC code on the package of the absorbent article. For example, the provision of this information may be motivated by binding the scanning of the UPC code to the reward program. In some embodiments, the user provides information about the number of disposable articles they change per day. In some embodiments, a statistical model is utilized regarding the number of absorbent articles used by the wearer during the day based on age. Further, while some consumption predictions are schematically shown as the number of absorbent articles consumed, these consumption predictions may also be expressed in terms of the number of packages of absorbent articles. The consumption prediction from the package consumption representation may take into account the package configuration, as the number of absorbent articles contained within the package may vary based on absorbent article size, product lineup, and the like.

The recommendations provided by the disposable recommendation computing system may also include a value calculation for the user that incorporates various recommendation elements (such as the number of disposable articles to be consumed, the number of trips to the store, and other consumption metrics) to provide the user with an optimal overall value. With respect to automated shipment of products, recommendations may include a shipping schedule for various products in a product lineup. The disposable article recommendation computing system may be configured to receive feedback from the user to provide information regarding the manner in which the recommendation was received. The feedback can be used to improve or otherwise refine the particular recommendation by collecting feedback across multiple users and/or improve the quality of the overall model over time. Further, based on the disposable consumption prediction, the disposable recommendation computing system may schedule a notification to the user when it is time to purchase additional disposable. Such notifications may be in any suitable format and may be scheduled over any suitable communication medium. In some embodiments, the notification is a text message, an email message, an in-application message, a calendar appointment, a reminder, a smart speaker notification, a pop-up notification, a geo-fence notification (e.g., when physically near a store), or a combination thereof. In addition, the user may also be provided with a subscription purchasing program, such that the disposable articles are routinely sent to the user, and the recommended size of the disposable articles automatically increases over time. In some configurations, a bulk shipment is provided that contains quantities of all of the various sizes of disposable articles predicted to be used by the wearer over time.

Fig. 16 depicts another exemplary disposable article consumption prediction 1510 that may be displayed on an interface 1500 of a computing device 1520. The disposable consumption forecast 1510 includes a recommended size 1506, a time period forecast 1514, and a quantity forecast 1516. In this embodiment, the time period prediction 1514 is expressed in terms of a date, and the quantity prediction 1516 comprises a rate of use. For example, usage rates may be adjusted as input such that associated recommendations change with user-provided input regarding rate.

17-19 depict exemplary indications of purchase paths that may be provided to a user via an interface. The purchase path may be represented in any suitable format. Referring first to FIG. 17, an exemplary interface 1500 of a computing device 1720 illustrates a first example of a recommendation size 1706 and a purchase path 1708. In this embodiment, the purchase path 1708 is a retailer map that provides real-time location information based on GPS data provided by the computing device 1720. Accordingly, the purchase path 1708 may direct the user to the retailer's aisle to purchase the recommended size 1706 of disposable articles, and in some embodiments, provide additional information to the user, such as purchasable quantity, pricing data, and the like. FIG. 18 depicts an exemplary interface 1800 of a computing device 1820 showing another exemplary purchase path 1808 and a recommendation size 1806. The purchase path 1808 may include one or more web-based links to purchase a disposable article of the recommended size 1806 via an online retailer or an online marketplace forum. Fig. 19 depicts an exemplary interface 1900 of a computing device 1920 showing another exemplary purchase path 1908 and a recommended size 1906. The purchase path 1908 in this embodiment can include a map that includes driving directions to direct the user to a retailer to purchase the disposable article. As shown, additional purchase-related information may be provided, such as aisle location, purchasable quantity, pricing data, and the like.

Referring now to FIG. 20, exemplary simplified interfaces 2000A-C of a computing device 2020 are depicted. The interface 2000A schematically depicts a collection of various inputs from a user that may be utilized by the disposable article recommendation computing system. For example, providing the development token 2002 may assist the disposable recommendation computing system in recommending an appropriate product line. Exemplary developmental cues 2002 may include, but are not limited to, crawling, climbing on furniture, walking, starting toilet training, sleeping intuition, and the like. Additionally or alternatively, user-supplied input 2004 may be provided to a disposable item recommendation computing system to facilitate generating recommendations. Exemplary user-supplied inputs 2004 may include, for example, weight, length, head size, date of birth, gestational age of the subject at birth, and the like. Additional input may include geographic location, as certain areas have different desires with respect to the fit of the disposable article. For example, the geographic location may be ascertained based on the physical location of the user's computer device, or the geographic location may be provided by the user as part of the enrollment process. One or more inputs may be visualized and tracked separately or together with a population growth curve, such as a child's conventional CDC growth curve. Interface 2000B schematically depicts a collection of images of a subject. Similar to fig. 13, the interface 2000B may have an exemplary image preview pane 2006. The image preview pane 2006 may be used to properly align the subject in the field of view. The interactive element 2008 may be activated by a user to take a picture or series of pictures of the subject. Images collected by computing device 2020 may be processed in accordance with the present disclosure and interface 2000C may be presented to a user. The interface 2000C may present the recommendation 2010 in any suitable format. The recommendation 2010 may be based on both image analysis as described herein and additional information provided by the user via the interface 2000A. It should be appreciated that the recommendations 2010 may include a recommended product type, a recommended product lineup, and the like. The example interface 2000C also includes fit metrics 2012 to indicate a manner in which the recommended disposable article is expected to fit the subject.

Fig. 21 depicts an exemplary flowchart 2100 of a method for recommending a pre-made disposable article for a subject. At 2102, the method includes storing a plurality of disposable article sizing models corresponding to a respective plurality of pre-made disposable articles available for purchase. The sizing model may be stored by a disposable article recommendation computing system, as shown in fig. 1-6. In this regard, the sizing model may be stored in a centralized repository or may be stored locally at the end-user computer device. At 2104, images collected by at least one camera are received. The image may include a representation of the subject. The subject is a consumer of pre-formed disposable articles, such as an infant, baby, or toddler. The received image may be, for example, a still image, a collection of still images, or a video.

At 2106, a scale of the image that correlates a size of the representation of the subject in the image to a physical size of the subject is determined. The scale may be determined by any suitable technique, such as using a scale object in the image or inferring the scale based on known body part dimensions of the subject in the image (such as the height of the subject, the intra-pupillary distance or the head circumference of the subject). At 2108, physical properties of a representation of the subject in the image are determined. As described above, the physical attributes may include a plurality of joint positions of the subject, a distance between the detected aspects of the representation of the subject in the image, and an area of the subject or other quantifiable measurement of an aspect of the representation of the subject in the image.

At 2110, a plurality of fit parameters of the subject are determined based on the scale of the image and the physical attributes of the representation of the subject. Fit parameters may include, for example, an estimated waist circumference of the subject, an estimated thigh circumference of the subject, and an estimated rise measurement of the subject. At 2112, the plurality of fit parameters are applied to one or more of the plurality of disposable article sizing models. For example, it may be determined which sizing model captures each of the subject's estimated waist circumference, the subject's estimated thigh circumference, and the subject's estimated rise measurement.

At 2114, a recommended pre-made disposable article for the subject is determined based on application of the plurality of fit parameters to one or more of the plurality of disposable article sizing models. The recommended prefabricated disposable article is selected from a plurality of prefabricated disposable articles available for purchase. At 2116, an indication is provided to the user of a recommended pre-made disposable article for the subject.

Combination of

A. A computer-based method, the method comprising:

storing, by a disposable article recommendation computing system, a plurality of disposable article sizing models in a data store, the plurality of disposable article sizing models corresponding to a respective plurality of pre-manufactured disposable articles available for purchase;

receiving, by the disposable article recommendation computing system, an image collected by at least one camera, wherein the image comprises a representation of a subject, and the subject is a consumer of a pre-made disposable article;

determining, by the disposable recommendation computing system, a scale of the image that correlates a size of the representation of the subject in the image to a physical size of the subject;

determining, by the disposable recommendation computing system, a physical attribute of the representation of the subject in the image;

determining, by the disposable article recommendation computing system, fit parameters of the subject based on the scale of the image and the physical attributes of the representation of the subject;

applying, by the disposable article recommendation computing system, the plurality of fit parameters to one or more of the plurality of disposable article sizing models;

determining, by the disposable article recommendation computing system, a recommended pre-made disposable article for the subject based on application of the plurality of fit parameters to the one or more of the plurality of disposable article sizing models, wherein the recommended pre-made disposable article is selected from the plurality of pre-made disposable articles available for purchase; and is

Providing, by the disposable article recommendation computing system, an indication of the recommended pre-made disposable article for the subject.

B. The computer-based method according to paragraph a, wherein the recommended pre-made disposable of the subject includes any one of a recommended standard size pre-made disposable, a recommended product line pre-made disposable, and a recommended style of pre-made disposable.

C. The computer-based method according to any of paragraphs a-B, wherein the image includes a scale object positioned proximate to the representation of the subject.

D. A computer-based method according to paragraph C, wherein the physical dimensions of the scale object in the image are known to the disposable recommendation computing system.

E. The computer-based method according to any of paragraphs a-D, wherein determining the proportion of the image comprises: and determining the X-direction proportion and the Y-direction proportion.

F. The computer-based method of any of paragraphs a through E, further comprising:

prior to determining the plurality of fit parameters, processing, by the disposable article recommendation computing system, the image to account for a viewing angle between the subject and the at least one camera.

G. A computer-based method according to any of paragraphs a to F, wherein the physical attributes comprise joint positions of the subject.

H. The computer-based method of paragraph G, wherein the plurality of joint positions of the subject are determined based at least in part on a machine learning model.

I. A computer-based method according to any of paragraphs A to H, wherein the physical attributes comprise distances between the detected aspects of the representation of the subject in the image.

J. A computer-based method according to any of paragraphs A to I, wherein the physical attributes comprise any of a length, width, area and volume of an aspect of the representation of the subject in the image.

K. The computer-based method according to any one of paragraphs a-J, wherein the plurality of fit parameters are based on correlations with one or more of the physical attributes of the subject.

L. the computer-based method of any of paragraphs a-K, wherein the physical attributes include any of head width, inter-eye distance, torso length, hip width, and shoulder width.

M. the computer-based method according to any of paragraphs a-L, wherein the fit parameters include any of an estimated waist circumference of the subject, an estimated thigh circumference of the subject, and an estimated rise measurement of the subject.

N. the computer-based method of paragraph M, wherein each disposable article sizing model of the plurality of disposable article sizing models includes a waist circumference range, a thigh circumference range, and a rise measurement range.

O. the computer-based method of paragraph N, wherein applying the plurality of fit parameters to one or more of the plurality of disposable article sizing models comprises: modeling for each respective disposable article size to determine: whether the estimated waist circumference is within the waist circumference range, whether the estimated thigh circumference is within the thigh circumference range, and whether the estimated rise measurement is within the rise measurement range.

P. the computer-based method according to paragraph O, wherein the recommended prefabricated disposable of the subject is a prefabricated disposable of a plurality of different prefabricated disposables determined to be sized for the subject.

Q. the computer-based method of any of paragraphs a through P, further comprising:

receiving, by the disposable article recommendation computing system, one or more user supplied values.

R. the computer-based method according to paragraph Q, wherein the user-supplied values include any of the age of the subject, the weight of the subject, the race of the subject, the sex of the subject, the height of the subject, the gestational age of the subject at birth, and the head circumference of the subject.

S. the computer-based method of any of paragraphs a through R, further comprising:

determining, by the disposable article recommendation computing system, a disposable article consumption prediction; and is

Providing, by the disposable recommendation computing system, the disposable consumption prediction.

T. the computer-based method of paragraph S, wherein the disposable consumption prediction is based on one or more of the user-supplied values.

U. the computer-based method of paragraph T, wherein the disposable consumption prediction is based on one or more of the user-supplied values and one or more of the determined physical attributes.

V. the computer-based method according to any of paragraphs R-U, wherein the disposable consumption prediction comprises an estimated number of the recommended pre-made disposables to be used by the subject.

W. the computer-based method of any of paragraphs R-V, wherein the disposable consumption prediction is associated with any of a product, a product size, and a product lineup.

X. the computer-based method of any of paragraphs R-W, wherein the disposable consumption prediction comprises an estimated amount of time that the recommended pre-made disposable will fit the subject.

Y. the computer-based method of any of paragraphs R through X, further comprising:

sending, by the disposable article recommendation computing system, a purchase reminder notification to a remote computing device based on the disposable article consumption prediction.

Z. the computer-based method of any of paragraphs R-Y, further comprising:

registering, by the disposable recommendation computing system, the user into a subscription purchasing program for pre-made disposable items.

A computer-based method according to any of paragraphs R to Z, further comprising:

the recommended size of these recommended pre-made disposable articles is automatically increased after a certain period of time.

The computer-based method of any of paragraphs a through AA, wherein the image is one or more still images of the subject collected by a mobile computing device.

Ac. the computer-based method of paragraph AB, wherein the one or more still images of the subject are collected by a backward camera of the mobile computing device.

AD. the computer-based method according to any of paragraphs a to AC, the method further comprising:

receiving, by the disposable recommendation computing system, a plurality of images of the subject collected over a period of time, wherein the period of time is greater than three days; and is

Predicting, by the disposable article recommendation computing system, a disposable article consumption prediction based on the growth pattern of the subject determined from the plurality of images.

AE. the computer-based method according to paragraph AD, the method further comprising:

recommending, by the disposable recommendation computing system, a prediction of pre-made disposable size change over time based on the growth pattern prediction of the subject determined from the plurality of images.

A computer-based method according to any of paragraphs a to AE, further comprising:

at least one image collection guidance tool is provided by the disposable article recommendation computing system.

The computer-based method of paragraph AF, wherein the at least one image collection guidance tool comprises a graphic overlay for presentation to a user during image collection.

AH. the computer-based method according to paragraph AF, wherein the at least one image collection guidance tool comprises a proportional object graph guidance area.

Ai the computer-based method of any of paragraphs a through AH, wherein processing the image to determine the scale of the image comprises: using the known body part size of the subject.

The computer-based method according to paragraph AI, wherein the known body part size of the subject is any one of the height of the subject, the intra-pupillary distance of the subject, and the head circumference of the subject.

AK. the computer-based method according to any of paragraphs a to AJ, wherein processing the image to determine the scale of the image includes: a plurality of images of the subject are utilized in a photogrammetry process.

AL. the computer-based method according to any of paragraphs a to AK, wherein processing the image to determine the scale of the image comprises: the deformation of the structured light projected onto the subject is analyzed.

AM. the computer-based method according to any of paragraphs a to AL, the method further comprising:

determining, by the disposable recommendation computing system, a geographic location of the subject, wherein the recommended pre-made disposable of the subject is based on the geographic location.

AN. the computer-based method according to any of paragraphs a-AM, wherein the image is a top view of the whole body of the subject, and wherein the subject is any one of an infant, a baby, and a toddler.

AO. the computer-based method according to paragraph AN, wherein the subject is supine in the image.

AP. the computer-based method according to paragraph AO, the method further comprising:

verifying, by an error checking module of the disposable article recommendation computing system, one or more verification parameters of the image.

AQ. the computer-based method according to paragraph AP, wherein verifying the verification parameters includes: verifying any of a presence of the infant, a posture of the infant, a presence of the scale object, an orientation and a perspective of the camera relative to the scale object, and an orientation and a perspective of the camera relative to the subject.

Ar. the computer-based method according to paragraph AO, wherein verifying the verification parameters comprises: any of the physical attributes of the scale object are verified.

AS. the computer-based method according to paragraph AR, the method further comprising: the image is downsampled prior to verifying the physical properties of the scale object.

AT. the computer-based method according to any of paragraphs a to AS, the method further comprising:

providing, by the disposable recommendation computing system, an indication of a purchase path of the recommended pre-made disposable for the subject.

AU. the computer-based method of any of paragraphs a-AT, wherein the prefabricated disposable article is any one of a disposable diaper and a disposable training pant.

AV. the computer-based method of any one of paragraphs a to AU, wherein the image collected by the at least one camera is a single image.

AW. A computer-based system, the system comprising:

a data store, wherein a plurality of disposable article sizing models are stored by the data store, the plurality of disposable article sizing models corresponding to pre-made disposable articles of respective sizes available for purchase; and is

A disposable article recommendation computing system comprising a computer-readable medium having computer-executable instructions stored thereon configured to direct one or more computer processors to:

receiving an image of a subject collected by a remote mobile computing device;

determining a scale of the image;

processing the image to determine a physical attribute of the object;

determining fit parameters of the subject based on the scale of the image and the physical attributes of the subject;

comparing the plurality of fit parameters to one or more of the plurality of disposable article sizing models;

determining a recommended pre-made disposable article for the subject based on the comparison of the plurality of fit parameters to the one or more of the plurality of disposable article sizing models; and is

An indication of the recommended pre-made disposable article is sent to the remote mobile computing device.

AX. the computer-based system according to paragraph AW, wherein the recommended pre-made disposable of the subject includes any one of a recommended standard size pre-made disposable, a recommended product line of pre-made disposable, and a recommended style of pre-made disposable.

AY., the image including a scale object positioned proximate to the representation of the subject according to any of paragraphs AW-AX.

AZ. the computer-based system according to paragraph AY, wherein the physical dimensions of the scale object in the image are known to the disposable article recommendation computing system.

The computer-based system according to paragraph AZ, wherein the physical dimensions of the proportional object in the image known to the computing system for the disposable article recommendation include any of a corner-to-corner dimension, a height dimension, a width dimension, and a radius dimension.

BB., the computer-based system of any one of paragraphs AW to BA, wherein the computer-executable instructions are further configured to instruct one or more computer processors to:

prior to determining the plurality of fit parameters, processing the image to account for a viewing angle between the subject and the at least one camera.

The computer-based system according to any of paragraphs AW-BB, wherein the physical attributes comprise joint positions of the subject.

BD. the computer-based system according to paragraph BC, wherein the plurality of joint positions of the subject are determined based at least in part on a machine learning model.

BE. the computer-based system according to any of paragraphs AW to BD, wherein the physical attributes include distances between the detected aspects of the representation of the subject in the image.

BF. the computer-based system according to any one of paragraphs AW to BE, wherein the physical attributes include any one of a circumference, an area and a volume of an aspect of the representation of the subject in the image.

BG., the computer-based system according to any one of paragraphs AW to BF, wherein the plurality of fit parameters are based on correlations with one or more of the physical attributes of the subject.

BH. the computer-based system of any one of paragraphs AW-BG, wherein the physical attributes comprise any one of head width, eye-to-eye distance, torso length, hip width, and shoulder width.

BI. the computer-based system of any one of paragraphs AW-BH, wherein the fit parameters include any one of an estimated waist circumference of the subject, an estimated thigh circumference of the subject, and an estimated rise measurement of the subject.

BJ. the computer-based system of paragraph BI wherein each disposable article sizing model of the plurality of disposable article sizing models includes a waist circumference range, a thigh circumference range, and a rise measurement range.

The computer-based system of paragraph BJ, wherein the comparison of the plurality of fit parameters to one or more of the plurality of disposable article sizing models comprises: modeling for each respective disposable article size to determine: whether the estimated waist circumference is within the waist circumference range, whether the estimated thigh circumference is within the thigh circumference range, and whether the estimated rise measurement is within the rise measurement range.

BL. the computer-based system of paragraph BK wherein the recommended prefabricated disposable article of the subject is a prefabricated disposable article of a plurality of different prefabricated disposable articles determined to be sized for the subject.

A computer-based system according to any of paragraphs AW to BL, wherein the computer-executable instructions are further configured to instruct one or more computer processors to:

one or more user-supplied values are received, as entered into the remote computing device.

BN. the computer-based system of paragraph BM wherein the user-supplied values include any one of the age of the subject, the weight of the subject, the race of the subject, the sex of the subject, the height of the subject, the gestational age of the subject at birth, and the head circumference of the subject.

BO. the computer-based system according to any of paragraphs BM to BN, wherein the computer-executable instructions are further configured to instruct one or more computer processors to:

determining a disposable consumption prediction; and is

Providing the disposable consumption prediction to the remote mobile computing device.

Bp. the computer-based system according to paragraph BO, wherein the disposable consumption prediction is based on one or more of the user-supplied values.

The computer-based system of paragraph BP, wherein the disposable article consumption prediction is based on one or more of the user-supplied values and one or more of the determined physical attributes.

BR. the computer-based system according to any of paragraphs BO to BQ, wherein the disposable consumption prediction comprises an estimated number of the recommended pre-made disposables to be used by the subject.

BS. the computer-based system of any one of paragraphs BO to BR, wherein the disposable item consumption prediction is associated with any one of a product, a product size and a product lineup.

BT. the computer-based system of any of paragraphs BO to BS, wherein the disposable consumption prediction comprises an estimated amount of time that the recommended pre-made disposable will fit the subject.

BU. the computer-based system of any of paragraphs BO to BT, wherein the computer-executable instructions are further configured to instruct one or more computer processors to:

sending a purchase reminder notification to the remote computing device based on the prediction of disposable consumption.

BV., the computer-based system of any one of paragraphs BO to BU, wherein the computer-executable instructions are further configured to instruct one or more computer processors to:

the user is registered into a subscription purchasing program for pre-made disposable articles.

BW. the computer-based system according to any of paragraphs BO to BV, wherein the computer-executable instructions are further configured to instruct one or more computer processors to:

BX., the computer-based system according to any of paragraphs AW-BW, wherein the image is one or more still images of the subject collected by the remote mobile computing device.

BY. the computer-based system of paragraph BX, wherein the one or more still images of the subject are collected by a rear-facing camera of the remote mobile computing device.

BZ., the computer-based system of any of paragraphs AW through BY, wherein the computer-executable instructions are further configured to instruct the one or more computer processors to:

receiving a plurality of images of a subject collected over a period of time, wherein the period of time is greater than three days; and is

Predicting a disposable consumption prediction based on the determined growth pattern of the subject from the plurality of images.

CA. the computer-based system according to paragraph BZ, wherein the computer-executable instructions are further configured to instruct one or more computer processors to:

recommending a prediction of pre-made disposable article size change over time based on the growth pattern prediction of the subject determined from the plurality of images.

CB., the computer-based system of any one of paragraphs AW to CA, wherein the computer-executable instructions are further configured to instruct one or more computer processors to:

at least one image collection guidance tool is provided.

The computer-based system of paragraph CB, wherein the at least one image collection guidance tool includes a graphic overlay for presentation to a user during image collection.

The computer-based system of any of paragraphs AW through CC, wherein processing the image to determine the scale of the image comprises: using the known body part size of the subject.

The computer-based system of paragraph CD, wherein the known body part size of the subject is any one of the height of the subject and the head circumference of the subject.

The computer-based system of any of paragraphs AW through CE, wherein processing the image to determine the scale of the image comprises: a plurality of images of the subject are utilized in a photogrammetry process.

CG. the computer-based system of any of paragraphs AW through CF wherein processing the image to determine the scale of the image comprises: consider the vertical separation between the plane of the scale object and the plane in which the physical attribute of the subject lies.

CH. the computer-based system of any of paragraphs AW to CG wherein processing the image to determine the scale of the image comprises: the deformation of the structured light projected onto the subject is analyzed.

CI., the computer-based system of any of paragraphs AW through CH, wherein the computer-executable instructions are further configured to instruct one or more computer processors to:

determining a geographic location of the subject, wherein the recommended pre-made disposable article of the subject is based on the geographic location.

The computer-based system of any of paragraphs AW-CI, wherein the image is a whole-body top view of the subject, and wherein the subject is any of an infant, a baby, and a toddler.

The computer-based system according to paragraph CJ, wherein the subject is supine in the image.

CL. the computer-based system according to paragraph CK, wherein the computer-executable instructions are further configured to instruct one or more computer processors to:

one or more verification parameters for verifying the image.

CM. the computer-based system according to paragraph CL, wherein verifying the verification parameters includes: verifying any of a presence of the infant, a posture of the infant, a presence of the scale object, an orientation and a perspective of the camera relative to the scale object, and an orientation and a perspective of the camera relative to the subject.

CN., the computer-based system of any one of paragraphs AW to CM, wherein the computer-executable instructions are further configured to instruct one or more computer processors to:

providing an indication of a purchase path of the recommended pre-made disposable article for the subject.

Co. the computer-based system of any of paragraphs AW through CN, wherein the prefabricated disposable article is any of a disposable diaper and a disposable training pant.

A computer-based system according to any of paragraphs AW to CO, wherein the image received from the remote mobile computing device is a single image.

A computer-based method, the method comprising:

storing a plurality of disposable article sizing models corresponding to a respective plurality of pre-made disposable articles available for purchase;

receiving an image collected by at least one camera, wherein the image comprises a representation of a subject, and the subject is a consumer of a pre-made disposable article;

determining a scale of the image that correlates a size of the representation of the subject in the image with a physical size of the subject;

determining, by image processing, a physical property of the representation of the subject in the image;

determining fit parameters of the subject based on the scale of the image and the physical attributes of the representation of the subject;

The dimensions and values disclosed herein are not to be understood as being strictly limited to the exact numerical values recited. Rather, unless otherwise specified, each such dimension is intended to mean both the recited value and a functionally equivalent range surrounding that value. For example, a dimension disclosed as "40 mm" is intended to mean "about 40 mm".

Each document cited herein, including any cross referenced or related patent or patent application and any patent application or patent to which this application claims priority or its benefits, is hereby incorporated by reference in its entirety unless expressly excluded or otherwise limited. The citation of any document is not an admission that it is prior art with any disclosure of the invention or the claims herein or that it alone, or in combination with any one or more of the references, teaches, suggests or discloses any such invention. Further, to the extent that any meaning or definition of a term in this document conflicts with any meaning or definition of the same term in a document incorporated by reference, the meaning or definition assigned to that term in this document shall govern.

While particular embodiments of the present invention have been illustrated and described, it would be obvious to those skilled in the art that various other changes and modifications can be made without departing from the spirit and scope of the invention. It is therefore intended to cover in the appended claims all such changes and modifications that are within the scope of this invention.

Claims

1. A computer-based method, the method comprising:

determining, by the disposable article recommendation computing system, a scale of the image that correlates a size of the representation of the subject in the image with a physical size of the subject;

determining, by the disposable article recommendation computing system, a physical attribute of the representation of the subject in the image;

determining, by the disposable article recommendation computing system, a plurality of fit parameters for the subject based on the scale of the image and the physical attribute of the representation of the subject;

Providing, by the disposable article recommendation computing system, an indication of the recommended pre-made disposable articles for the subject.

2. The computer-based method of claim 1, wherein the recommended pre-made disposable of the subject includes any of a recommended standard size pre-made disposable, a recommended product line pre-made disposable, and a recommended style of pre-made disposable.

3. The computer-based method of claim 1, wherein the image includes a scale object positioned proximate to the representation of the subject.

4. The computer-based method according to any of the preceding claims, the method further comprising:

5. The computer-based method of any of the preceding claims, wherein the physical attributes include any of head width, eye separation distance, torso length, hip width, and shoulder width.

6. The computer-based method according to any of the preceding claims, wherein the fit parameter includes any of an estimated waist circumference of the subject, an estimated thigh circumference of the subject, and an estimated rise measurement of the subject.

7. The computer-based method according to claim 6, wherein each disposable article sizing model of the plurality of disposable article sizing models includes a waist circumference range, a thigh circumference range, and a rise measurement range.

8. The computer-based method of claim 7, wherein applying the plurality of fit parameters to one or more of the plurality of disposable article sizing models comprises: modeling for each respective disposable article size to determine: whether the estimated waist circumference is within the waist circumference range, whether the estimated thigh circumference is within the thigh circumference range, and whether the estimated rise measurement is within the rise measurement range.

9. The computer-based method of claim 8, wherein the recommended prefabricated disposable article of the subject is one of a plurality of different prefabricated disposable articles determined to be sized for the subject.

10. The computer-based method of claim 1, the method further comprising:

11. The computer-based method of claim 10, wherein the user-supplied value comprises any one of an age of the subject, a weight of the subject, an ethnicity of the subject, a gender of the subject, a height of the subject, a gestational age of the subject at birth, and a head circumference of the subject.

12. The computer-based method of claim 10, the method further comprising:

Providing, by the disposable article recommendation computing system, the disposable article consumption prediction.

13. The computer-based method of claim 12, wherein the disposable consumption prediction is based on one or more of the user-supplied values.

14. The computer-based method of claim 13, wherein the disposable consumption prediction is based on one or more of the user-supplied values and one or more of the determined physical attributes.

15. The computer-based method of claim 12, the method further comprising: