CN112232113A - Person identification method, person identification device, storage medium, and electronic apparatus - Google Patents
Person identification method, person identification device, storage medium, and electronic apparatus Download PDFInfo
- Publication number
- CN112232113A CN112232113A CN202010923634.XA CN202010923634A CN112232113A CN 112232113 A CN112232113 A CN 112232113A CN 202010923634 A CN202010923634 A CN 202010923634A CN 112232113 A CN112232113 A CN 112232113A
- Authority
- CN
- China
- Prior art keywords
- person
- image
- peer
- image set
- images
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Medical Informatics (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Collating Specific Patterns (AREA)
Abstract
The disclosure provides a person identification method, a person identification device, a storage medium and electronic equipment, and relates to the technical field of computers. The personnel identification method comprises the following steps: acquiring a first image set containing images of a person to be identified; determining a second image set comprising images of at least one first co-worker according to the first image set, wherein the first co-worker is a co-worker of the person to be identified; determining a third image set comprising images of at least one second peer person from the second image set, the second peer person being a peer person of the first peer person; and comparing the image of the person to be identified in the first image set with the image of the second person in the same row in the third image set to determine whether the person to be identified is the second person in the same row. The method and the device solve the problem that the identity of the person cannot be determined when the face is shielded or not clear, and improve the accuracy of person identification.
Description
Technical Field
The present disclosure relates to the field of computer technologies, and in particular, to a person identification method, a person identification apparatus, a computer-readable storage medium, and an electronic device.
Background
The face recognition needs to capture a clear and complete face image, and if the face is shielded, such as wearing a mask, a hat and the like to shield the face, or being shielded by surrounding people and obstacles, or capturing the shadow of a person, the identity of the person cannot be accurately recognized.
It is to be noted that the information disclosed in the above background section is only for enhancement of understanding of the background of the present disclosure, and thus may include information that does not constitute prior art that is already known to a person of ordinary skill in the art.
Disclosure of Invention
The disclosure provides a person identification method, a person identification device, a computer readable storage medium and an electronic device, thereby solving at least to a certain extent the problem that the identity of a person cannot be accurately identified when a face is blocked in the related art.
Additional features and advantages of the disclosure will be set forth in the detailed description which follows, or in part will be obvious from the description, or may be learned by practice of the disclosure.
According to a first aspect of the present disclosure, there is provided a person identification method comprising: acquiring a first image set containing images of a person to be recognized, wherein the images of the person to be recognized contain faces of the person to be recognized which do not meet a first quality condition or faces of the person to be recognized, and contain bodies of the person to be recognized which meet a second quality condition; determining a second image set comprising images of at least one first co-worker according to the first image set, wherein the first co-worker is a co-worker of the person to be identified; determining a third image set comprising images of at least one second peer person from the second image set, the second peer person being a peer person of the first peer person; wherein at least one image of the second peer in the third image set includes a face of the second peer satisfying the first quality condition and a body of the second peer satisfying the second quality condition; and comparing the image of the person to be identified in the first image set with the image of the second person in the same row in the third image set to determine whether the person to be identified is the second person in the same row.
In an alternative embodiment, the determining a second image set comprising images of at least one first peer from the first image set comprises: acquiring a second reference image meeting a first preset condition with images in the first image set, wherein the second image set comprises the second reference image; the first preset condition comprises any one or more of the following conditions: the difference of the shooting time is smaller than a first time threshold; the distance between the shooting places is smaller than a first distance threshold value; the scene similarity is greater than a first similarity threshold.
In an alternative embodiment, the determining a second image set including images of at least one first peer from the first image set further comprises: acquiring other images meeting a second preset condition with the second reference image, wherein the second image set also comprises the other images meeting the second preset condition; the second preset condition comprises any one or more of the following conditions: the second reference image and the other images meeting the second preset condition belong to a personnel image set of the same person, and the personnel image set is divided in advance; the similarity between the face image of the first person in the second reference image and the face images in the other images meeting the second preset condition is greater than a second similarity threshold.
In an alternative embodiment, the determining a third image set of images including images of at least one second fellow person from the second image set includes: acquiring a third reference image meeting a third preset condition with the images in the second image set, wherein the third image set comprises the third reference image; the third preset condition includes any one or more of: the difference of the shooting time is smaller than a third time threshold; the distance between the shooting places is smaller than a third distance threshold; the scene similarity is greater than a third similarity threshold.
In an alternative embodiment, the determining a third image set including images of at least one second fellow person from the second image set further comprises: acquiring other images meeting a fourth preset condition with the third reference image, wherein the third image set further comprises the other images meeting the fourth preset condition; the fourth preset condition includes any one or more of: the third reference image and the other images meeting the third preset condition belong to a personnel image set of the same person, and the personnel image set is divided in advance; the similarity between the face image of the second person in the third reference image and the face images in the other images under the third preset condition is greater than a fourth similarity threshold.
In an alternative embodiment, at least one first peer is determined from the second image set by: and in the second image set, carrying out clear detection on the face images of people except the people to be identified, and determining at least one person of which the clear detection result of the face image meets the first quality condition as the first person in parallel.
In an alternative embodiment, at least one second fellow person is determined from the third image set by: and in the third image set, performing clear detection on the face images and the human body images of people except the first peer, and determining at least one person whose face image clear detection result meets the first quality condition and whose human body image clear detection result meets the second quality condition as the second peer.
In an alternative embodiment, the comparing the image of the person to be identified in the first image set with the image of the second person in the third image set to determine whether the person to be identified is the second person in the same group includes: and when the similarity between the human body image of the person to be identified and the human body image of the second peer is detected to be greater than a fifth similarity threshold, determining that the person to be identified and the second peer are the same person.
In an alternative embodiment, when a plurality of first persons in a same row are determined from the second image set, and a corresponding at least one second person in a same row is determined for each first person in the third image set, the comparing the image of the person to be identified in the first image set with the image of the second person in the third image set to determine whether the person to be identified is the second person includes: determining the image of a second peer with the highest image similarity with the person to be identified as the image of the candidate person in second peers corresponding to each first peer respectively; clustering the images of the candidate persons to enable the candidate persons in each category to be the same person; and determining that the person to be identified and the candidate person in the category containing the largest number of images are the same person.
According to a second aspect of the present disclosure, there is provided a person identification method comprising: acquiring a first image set containing images of a person to be recognized, wherein the images of the person to be recognized contain faces of the person to be recognized which do not meet a first quality condition or faces of the person to be recognized, and contain bodies of the person to be recognized which meet a second quality condition; determining a second image set comprising images of at least one reference person from the first image set; wherein at least one image of the reference person in the second image set includes a human face of the reference person satisfying the first quality condition and a human body of the reference person satisfying the second quality condition, and the human body of the reference person in the at least one image of the reference person is similar to the human body of the person to be recognized in the image of the person to be recognized; acquiring at least one image of a first peer and at least one image of a second peer according to the first image set and/or the second image set, wherein the first peer is the peer of the person to be identified, and the second peer is the peer of the reference person; and comparing the image of the first peer with the image of the second peer to determine whether the person to be identified is the reference person.
According to a third aspect of the present disclosure, there is provided a person identification apparatus comprising: the image acquisition module is used for acquiring a first image set containing images of people to be recognized, wherein the images of the people to be recognized contain faces of the people to be recognized which do not meet a first quality condition or faces of the people to be recognized, and contain human bodies of the people to be recognized which meet a second quality condition; a first determining module, configured to determine, according to the first image set, a second image set including images of at least one first peer, where the first peer is a peer of the to-be-identified person; a second determining module, configured to determine, according to the second image set, a third image set including images of at least one second peer person, where the second peer person is a peer person of the first peer person; wherein at least one image of the second peer in the third image set includes a face of the second peer satisfying the first quality condition and a body of the second peer satisfying the second quality condition; and the person comparison module is used for comparing the image of the person to be identified in the first image set with the image of the second person in the same row in the third image set so as to determine whether the person to be identified is the second person in the same row.
According to a fourth aspect of the present disclosure, there is provided a person identification device comprising: the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring a first image set containing images of people to be recognized, and the images of the people to be recognized contain faces of the people to be recognized which do not meet a first quality condition or faces of the people to be recognized and contain bodies of the people to be recognized which meet a second quality condition; an image set determination module for determining a second image set comprising images of at least one reference person from the first image set; wherein at least one image of the reference person in the second image set includes a human face of the reference person satisfying the first quality condition and a human body of the reference person satisfying the second quality condition, and the human body of the reference person in the at least one image of the reference person is similar to the human body of the person to be recognized in the image of the person to be recognized; a second obtaining module, configured to obtain an image of at least one first peer and an image of at least one second peer according to the first image set and/or the second image set, where the first peer is a peer of the to-be-identified person, and the second peer is a peer of the reference person; and the personnel comparison module is used for comparing the image of the first peer with the image of the second peer so as to determine whether the personnel to be identified is the reference personnel.
According to a fifth aspect of the present disclosure, there is provided a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements any one of the above-mentioned person identification methods.
According to a sixth aspect of the present disclosure, there is provided an electronic device comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform any of the above-described person identification methods via execution of the executable instructions.
The technical scheme of the disclosure has the following beneficial effects:
and establishing the association between the person to be identified and the second person in the same row by using the person in the first image set, the second image set and the third image set as an intermediate object according to the person in the same row relationship, and further determining the identity of the person to be identified. On the one hand, the scheme is suitable for the condition that the face of the person to be recognized is shielded or unclear, the problem that the identity of the person cannot be determined when the face of the person is shielded or unclear in the related technology is solved, and the accuracy of person recognition is improved. On the other hand, the scheme can be realized on the basis of the first image set, the second image set and the third image set during monitoring or snapshot, additional information is not needed, the realization process is simple, and the practicability is high.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure. It is apparent that the drawings in the following description are only some embodiments of the present disclosure, and that other drawings can be obtained from those drawings without inventive effort for a person skilled in the art.
FIG. 1 illustrates a system architecture diagram of an environment in which the present exemplary embodiment operates;
FIG. 2 illustrates a flow chart of a person identification method in the present exemplary embodiment;
FIG. 3 shows a flowchart for comparing a person to be identified with a second peer in the exemplary embodiment;
FIG. 4 illustrates a flow chart of another method of person identification in the exemplary embodiment;
fig. 5 shows a flowchart of yet another person identification method in the present exemplary embodiment;
fig. 6 is a block diagram showing the structure of a person identifying apparatus in the present exemplary embodiment;
fig. 7 is a block diagram showing the structure of another person identification apparatus in the present exemplary embodiment;
fig. 8 shows an electronic device for implementing the above method in the present exemplary embodiment.
Detailed Description
Example embodiments will now be described more fully with reference to the accompanying drawings. Example embodiments may, however, be embodied in many different forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the disclosure. One skilled in the relevant art will recognize, however, that the subject matter of the present disclosure can be practiced without one or more of the specific details, or with other methods, components, devices, steps, and the like. In other instances, well-known technical solutions have not been shown or described in detail to avoid obscuring aspects of the present disclosure.
Furthermore, the drawings are merely schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus their repetitive description will be omitted. Some of the block diagrams shown in the figures are functional entities and do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor devices and/or microcontroller devices.
Herein, "first", "second", "third", etc. are labels for specific objects, and do not limit the number or order of the objects.
In the related art, when a face image is unclear or incomplete, human body features need to be recognized, however, the human body features are not unique, so that the identity of a person is recognized through the human body features, and the accuracy is low when a large-range human body comparison is performed.
In view of one or more of the above problems, exemplary embodiments of the present disclosure provide a person identification method. Fig. 1 shows a system architecture diagram of an environment in which the method operates, including a photographing apparatus 110, a recognition processing apparatus 120, and a database 130. The shooting device 110 may be a monitoring camera arranged in a street, a community, a station, or the like, and is used for capturing images of people; the recognition processing device 120 may be a background computer or server for performing a process of person recognition; the database 130 may be a person database, a certificate database, a personnel information database, etc. for providing standard information or reference information of the identity of a person.
It should be understood that the number of the devices in fig. 1 is merely exemplary, and any number of the photographing apparatus 110, the recognition processing apparatus 120, or the database 130 may be provided according to implementation needs.
Fig. 2 shows a schematic flow of a person identification method in the present exemplary embodiment, which may be executed by the above-described identification processing apparatus 120, including the following steps S210 to S240:
step S210, acquiring a first image set containing images of a person to be identified;
step S220, determining a second image set containing at least one image of a first co-worker according to the first image set, wherein the first co-worker is a co-worker of the person to be identified;
step S230, determining a third image set containing at least one image of a second person in the same row according to the second image set, wherein the second person in the same row is the person in the same row of the first person in the same row;
step S240, comparing the image of the person to be identified in the first image set with the image of the second fellow person in the third image set to determine whether the person to be identified is the second fellow person.
Based on the method, the relationship among the persons in the first image set, the second image set and the third image set is utilized, the first person in the same row is taken as an intermediate object, the association between the person to be identified and the second person in the same row is established, and the identity of the person to be identified is further determined. On the one hand, the scheme is suitable for the condition that the face of the person to be recognized is shielded or unclear, the problem that the identity of the person cannot be determined when the face of the person is shielded or unclear in the related technology is solved, and the accuracy of person recognition is improved. On the other hand, the scheme can be realized on the basis of the first image set, the second image set and the third image set during monitoring or snapshot, additional information is not needed, the realization process is simple, and the practicability is high.
Each step is described in detail below.
In step S210, a first set of images comprising images of a person to be identified is acquired.
The first image set may include one or more images, and the images of the first image set include the person to be identified. For example, a surveillance camera captures one or more photographs of a person entering the front of a building, i.e., a first set of images, and the person to be identified is identified. In the present exemplary embodiment, the image of the person to be recognized contains the face of the person to be recognized that does not satisfy the first quality condition or does not contain the face of the person to be recognized, and contains the human body of the person to be recognized that satisfies the second quality condition. The first quality condition refers to the requirement of the face recognition on the image quality, and the second quality condition refers to the requirement of the human body recognition on the image quality. In other words, in the first image set, the person to be recognized may be in a state where the face is blocked or unclear, or in a state where the face is not photographed (for example, only the back or the body part below the neck is photographed), the identity of the person cannot be determined directly through face recognition, but the image of the body part is in a clear state, and the human body characteristics of the person to be recognized may be recognized, for example, the person to be recognized may monitor the photographed image when wearing a mask.
In an alternative embodiment, partial images of the person to be identified can be extracted from the captured images over a certain period of time, forming a first image set. For example, when a monitoring person sees an abnormal person in a monitoring picture, the monitoring person can select a person to be identified in the current monitoring picture by clicking a person identification option, the system generates a rectangular detection frame, captures an image of the person to be identified to form a first reference image, and detects the person to be identified according to the first reference image and captures the person to be identified in all the monitoring pictures near the current monitoring picture for a period of time (such as 1 minute before and after the current monitoring picture) to form a first image set. Or, the human body detection model automatically detects whether a person exists in the video or the image, and cuts out a local image containing the person to be identified from the video frame of the person.
In step S220, a second image set is determined from the first image set, which second image set contains images of at least one first co-worker, which is a co-worker of the person to be identified.
Ideally, it is desirable to find a fellow person who has a social relationship with the person to be identified, i.e., a person who knows with the person to be identified and who fellow both, and it is undesirable to find a person who accidentally fellow without knowing with the person to be identified, because the person who knows with the person to be identified is likely to fellow with the person to be identified only once, whereas a person who knows with the person to be identified is more likely to accidentally fellow with the person to be identified and no other fellow occurs. If the person to be recognized shoots the unclear face in the same row of the person to be recognized and the person A acquainted with the person to be recognized, but the person to be recognized shoots the clear face in the other same row, the clear face and the unclear face of the person to be recognized can be associated through the person A. Whether people are in the same line or not can be judged only according to the spatio-temporal relation of the people according to the video and image information, so that the social relation is estimated according to the information of the same line (for example, A and B are in the same line for 3 times, and the two are considered to have the social relation). It is therefore understood that in the exemplary embodiment, the fellow persons do not necessarily refer to the fellow persons having social relationships, for example, two persons who are not acquainted with each other, walk into the front door of the building at the same time and go to the elevator hall, and the two persons form the fellow relationship at the time, and one person can be considered as the fellow person of the other person. A picture of a person at a certain time range and/or location range from the time, location, and/or location range of the image including the person to be identified can be considered as an image of a fellow person of the person to be identified, and as taken while the fellow person fell with the person to be identified. The second set of images contains these images. For example, the image of the person to be identified in the first image set is a monitoring image captured when the person to be identified enters a certain building at the front door, and the image of the person captured by the same camera at 13:00:05 is taken as the image in the second image set. In an alternative embodiment, partial images of the person can be extracted from images which are at a time range and/or a location range from the image recording time, location and/or location range of the person to be identified, forming the second image set. In general, the person appearing in the second image set may be considered to be a person having a peer relationship with the person to be identified, i.e. a first peer.
In an alternative embodiment, step S220 may include:
and acquiring a second reference image meeting a first preset condition with the images in the first image set, wherein the second image set comprises the second reference image.
When the first image set includes a plurality of images, a second reference image satisfying a first predetermined condition may be acquired for each image in the first image set. For example, if the first image set includes 3 images P1, P2, and P3, the images satisfying the first predetermined condition with P1 are P11 and P12, the images satisfying the first predetermined condition with P2 are P21, and the images not satisfying the first predetermined condition with P3 are P11, P12, and P21, the second reference image is a reference image. Of course, the second reference image satisfying the first preset condition may be acquired only for a specific one or more images in the first image set.
The first preset condition may include any one or more of the following:
the difference of the shooting time is smaller than a first time threshold value. The first time threshold is a condition to be satisfied by the shooting time of the two images when the person is judged to be in the same line in the two images. In one example, the first preset condition is that the difference between the shooting times is smaller than a first time threshold and the shooting locations are the same. At this time, if the difference between the two image capturing times is within the first time threshold range, the persons in the two images are considered to be in the same line, and if the difference is beyond the range, the persons in the two images are not considered to be in the same line. The first time threshold may be determined according to daily living habits, for example, the time when two persons in a fellow companion pass through one camera in sequence is usually within 5s, and then the first time threshold may be set to 5 s. It will be appreciated that if the two images are taken at the same time, the difference between the times at which the two images are taken must be less than the first time threshold.
② the distance between the shooting places is smaller than the first distance threshold. The shooting location refers to a location where a person appears in the shot image. In general, the camera is located at the same place where the person appears in the image shot by the camera, so the shooting place can be replaced by the place where the camera is located. The first distance threshold is a condition to be satisfied by the shooting locations of the two images when the persons in the two images are judged to be in the same line. In one example, the first preset condition is that the difference between the shooting locations is smaller than a first location threshold and the shooting times are the same. At this time, if the difference between the two image capturing locations is within the first location threshold range, the persons in the two images are considered to be the same line, and if the difference is beyond the range, the persons in the two images are not considered to be the same line. The first distance threshold may be determined according to daily living habits, for example, the distance between two persons in a fellow row is usually within 5m, and may be simultaneously captured by cameras within 5m, and then the first distance threshold may be set to 5 m. It will be appreciated that if the two images are taken by the same camera, it is assumed that the distance between the two image taking locations is necessarily less than the first distance threshold. Whether the shooting places of the two images are the same can be judged through the ID of the camera, and the scene similarity of the images can be used for measurement. If the scenes contained in the two images are the same and only the individual persons in the images are different, the two images are considered to be shot at the same position, and the scene similarity of the two images is larger than a first similarity threshold value at the moment. It can be understood that, according to the scene similarity of the images, whether the shooting places of the two images are the same or not is measured, the image for comparing the scene similarity needs to be a panoramic image, rather than a person image obtained by picking up a person from the panoramic image.
And the difference of the appearance time of the mobile equipment information related to the image and the distance between the places are in a preset range. Some monitoring devices may detect information such as a Media Access Control (MAC) address, an International Mobile Subscriber Identity (IMSI), and the like of a Mobile device carried by a person who has passed, and transmit the information of the Mobile device to the identification processing device together with a captured image. Thus, the identification processing device can obtain the image and the mobile device information associated with the image, wherein the mobile device information comprises the time and place information of the mobile device detected by the monitoring device. When the difference between the detected time of the mobile equipment information related to a certain image and the detected time of the mobile equipment information related to the images in the first image set is within a preset time range and the distance between the detected points is within a preset distance range, the persons in the two images are considered to be in the same row.
And fourthly, the scene similarity is greater than a first similarity threshold. The scene similarity refers to extracting background parts from the two images respectively and calculating the similarity. The first similarity threshold is a condition to be met by judging that the backgrounds in the two images are the same; if 70% is taken as a first similarity threshold, respectively extracting background parts of the two images through a convolutional neural network to form two sub-images, calculating the similarity of the two sub-images, and determining the two sub-images as the same scene location if the similarity is more than 70%.
In practical application, the conditions (i) to (iv) may be arbitrarily combined and parameters in each condition may be set according to requirements, for example, the first preset condition may include both the conditions (i) and (ii), and the first time threshold and the first distance threshold are extremely small, and the second reference image is an image captured at the same time and the same place as the first image set image, that is, only an image of a person in the same frame as the person to be identified included in the first image set image is regarded as the second reference image. Thus, the person image having the same imaging time and the same imaging camera ID as those of the images in the first image set can be searched as the second reference image. Or, a panorama corresponding to each image in the first image set may be searched, and if the same panorama includes images of other persons, images of other persons in the panorama are cut out as the second reference image.
The present exemplary embodiment can form the second image set with the second reference image. In addition, in an optional implementation, step S220 may further include:
and acquiring other images meeting a second preset condition with the second reference image, wherein the second image set can also comprise the other images meeting the second preset condition.
That is, the image processing apparatus may perform expansion based on the second reference image, separately search for other images that meet the requirements, and form the second image set together with the second reference image.
The second preset condition may include any one or more of:
and the second reference image and other images meeting the second preset condition belong to the same person image set. In processing the monitoring images, images belonging to the same person may be previously divided into the same image set. When images are classified into image sets, images in which the degree of similarity between faces of persons included in a plurality of images is greater than a threshold value may be classified into the same image set, or a plurality of images in which persons included in a plurality of images can be determined to be the same person by other information such as mobile device information may be classified into the same image set. If the second reference image includes the first person a who is a person, it can be considered that the first person a is included in the other images when the reference image and the other images belong to the same person image set.
And sixthly, the similarity between the face image of the first peer in the second reference image and the face images in other images meeting a second preset condition is larger than a second similarity threshold. The second similarity threshold is used for judging whether the two face images are the same person or not. For example, the face image of the first person in the second reference image is used to perform face retrieval in other images, and when there are other images whose face similarity is greater than the second similarity threshold, the other images may be added to the second image set. If the second reference image includes the first person a in common, when the similarity between the second reference image and the face of the other image is greater than the second similarity threshold, it may be determined that the other image also includes the first person a in common. It is understood that when this condition is used as the second condition, the quality of the face in the second reference image and the other images is required to be such that it can be used for face comparison.
After other images meeting the second preset condition are found, the other images and the second reference image form a second image set together, so that the distribution of the second image set in time and space is widened, the range of subsequently acquiring a third image set is increased, and the probability that the third image set contains the images of the people to be identified is increased. For example, if the image of the person to be identified is P1, the panorama where P1 is located is F1, and F1 further includes an image P11 of another person, P11 is a second reference image. If the person image set in which P11 is located is found to be a person image set of zhang san among the previously divided person image sets, and the person image set includes P11 taken at the place 1 on the 1 st/1 day and P12 taken at the place 2 on the 1 st/3 day, P11 and P12 are taken as the second image set. Subsequently, a third image set can be searched according to the second image set, and whether the third image set contains other images of the person to be identified or not can be checked. And the persons contained in the images of the third image set are the persons in the same row as the persons contained in the images of the second image set. If the P12 is not found through the P11, the P11 and the P12 are used as the second image set, and the P11 is used as the second image set, when the third image set is subsequently found according to the second image set, only the P1 is probably found, and at this time, the purpose of identifying the person to be identified cannot be achieved.
Generally, the person appearing in the second image set is taken as the first peer. In practical application, a person with high quality can be selected as a first peer person to improve the accuracy of subsequent processing, wherein the main factor for evaluating the quality is whether the face image is clear or not. In particular, in an alternative embodiment, at least one first peer may be determined from the second image set by:
and in the second image set, carrying out clear detection on the face images of people except the people to be recognized, and determining at least one person of which the clear detection result of the face image meets a first quality condition as a first peer.
During the sharpness detection, if the second image set contains the images of the persons to be identified, the persons to be identified can be filtered from the second image set, the face areas are detected for the rest persons, each face area forms an independent face image, and a smaller face area can be filtered in the process; carrying out preprocessing such as denoising on each human face image; and calculating the image gradient, for example, calculating through a Tenengrad gradient function, a Brenner gradient function, a Laplacian gradient function and the like, and determining that the clear detection result meets the first quality condition when the gradient reaches a threshold corresponding to the first quality condition.
If the face images of a plurality of persons meet the first quality condition, the person with the better quality can be further determined as the first person in the following mode:
the other person who is closer to the person to be identified is preferentially determined as the first person in parallel. Specifically, the position association degree of the other people and the people to be identified can be measured by the number of people at intervals, wherein the number of people at intervals is 0 to indicate that the people to be identified are adjacent to the people to be identified, such as the people at the left and right sides of the people to be identified, the number of people at intervals is 1 to indicate that the people to be identified are still separated from the people to be identified, and the like. And preferentially determining other people with small space number as the first person in the same race.
Other persons who appear more frequently in the second image set are preferentially determined as the first fellow person. Specifically, the occurrence number of other persons can be measured by the occurrence ratio, for example, if some other person B appears in M1 images in the second image set, and the second image set has M0 images, the occurrence ratio of B is M1/M0. The person with a high occurrence ratio is preferably identified as the first person in the same group.
In practical application, the above different manners may be combined, for example, the number of the selected persons in the second image set is less than 2, the appearance ratio is greater than 70%, and the other persons whose face image sharpness detection result meets the first quality condition are the first peer persons.
In step S230, a third image set is determined from the second image set, which third image set contains images of at least one second fellow person, which second fellow person is a fellow person of the first fellow person.
In the present exemplary embodiment, the image of at least one second fellow person in the third image set includes a face of the second fellow person satisfying the first quality condition and a human body of the second fellow person satisfying the second quality condition. In other words, in the third image set, at least one image of the second peer includes a clear face part and a clear body part, so that face recognition and body recognition can be simultaneously realized.
A person picture that is distant from the image capturing time, location, and/or location range including the first fellow person by a certain time range and/or location range may be considered as an image of a second fellow person who is a first fellow person and is captured while the second fellow person is fellow the first fellow person, and these images may be referred to as a third reference image. The third image set includes a third reference image. In an alternative embodiment, a partial image of the person can be extracted as the third reference image from an image of the person at a time range and/or a location range from the image of the person including the first peer. In general, the person appearing in the third reference image may be a person having a peer relationship with the first peer, that is, a second peer.
In an alternative embodiment, step S230 may include:
and acquiring a third reference image meeting a third preset condition with the images in the second image set, wherein the third image set comprises the third reference image.
Similar to step S220, a third reference image that satisfies the third preset condition may be obtained for any one or more images in the second image set, and certainly, each image in the second image set may also be used to search for a third reference image that satisfies the third preset condition.
The third preset condition may include any one or more of:
and the difference of the shooting time is smaller than a third time threshold value. The third time threshold may refer to another condition that the shooting time of the two images is satisfied when the people in the two images are judged to be in the same line, and may be the same as or different from the first time threshold, for example, the time range of the images may be appropriately expanded, so that the third time threshold is greater than the first time threshold.
And the distance between the shooting places is smaller than a third distance threshold value. The third distance threshold may refer to another condition that the shooting locations of the two images are required to satisfy when the people in the two images are judged to be in the same row, and may be the same as or different from the first distance threshold, for example, the spatial range of the images may be appropriately expanded, so that the third distance threshold is greater than the first distance threshold.
And ninthly, the difference of the appearance time of the mobile equipment information related to the image and the distance between the places are in a preset range. Reference may be made to the above condition (c).
The likelihood of the scene in r is greater than a third likelihood threshold. With reference to the above condition, the third similarity threshold is another condition to be satisfied when the backgrounds of the two images are judged to be the same as the background of the first image, and may be the same as or different from the first similarity threshold.
In practical applications, the above conditions c to r may be arbitrarily combined and parameters in each condition may be set according to requirements, which is not limited by the present disclosure.
The present exemplary embodiment can form the third image set with the third reference image. In addition, in an optional implementation, step S230 may further include:
and acquiring other images meeting a fourth preset condition with the third reference image, wherein the third image set can also comprise the other images meeting the fourth preset condition.
That is, the image processing apparatus may perform expansion based on the third reference image, separately search for other images that meet the requirements, and form the third image set together with the third reference image.
The fourth preset condition may include any one or more of:
the third reference image and the other images satisfying the fourth preset condition belong to the same person's image set. The above condition (c) can be referred to.
The similarity between the face image of the second person in the third reference image and the face images in other images meeting a fourth preset condition is greater than a fourth similarity threshold. Referring to the above condition, the fourth similarity threshold is used to measure whether the two face images are the same person, and may be the same as or different from the second similarity threshold.
After other images meeting the fourth preset condition are found, the other images and the third reference image form a third image set together, so that the distribution of the third image set in time and space is widened, and the range of finding second persons in the same row is increased.
By the method, the second image set and the third image set can be ensured to have certain association in the aspects of time, space, scene and the like, so that the same-row relationship among people is ensured.
In an alternative embodiment, the second image set and the third image set may be the same image set, i.e., the second image set is directly used as the third image set.
Generally, each person appearing in the third image set can be regarded as a second person in the same row due to the strong correlation between the second image set and the third image set. In practical application, a person with high quality can be selected as a second peer person to improve the accuracy of subsequent processing, wherein the main factor for evaluating the quality is whether the face image and the human body image are clear or not. In particular, in an alternative embodiment, at least one second peer person may be determined from the third image set by:
and in the third image set, carrying out clear detection on the face images and the human body images of people except the first peer, and determining at least one person of which the clear detection result of the face image meets the first quality condition and the clear detection result of the human body image meets the second quality condition as a second peer.
When the definition detection is carried out, a first person who is a same person can be filtered from the third image set, the face regions and the human body regions can be respectively detected for the rest persons, each face region forms an independent face image, each human body region forms an independent human body image, the face images and the human body images are paired to form an image pair, and the face images and the human body images in each image pair belong to the same person; in the process, smaller face regions or human body regions can be screened; carrying out preprocessing such as denoising on each human face image and each human body image; calculating the image gradient, for example, calculating through a Tenengrad gradient function, a Brenner gradient function, a Laplacian gradient function, and the like, if the gradient of the face image in each image pair reaches a threshold corresponding to the first quality condition and the gradient of the human body image reaches a threshold corresponding to the second quality condition (generally, the threshold corresponding to the first quality condition is higher than the threshold corresponding to the second quality condition), it is determined that the image of the person simultaneously satisfies the first quality condition and the second quality condition, and the person can be used as a second peer.
If a plurality of persons meet the above condition, the superior person can be further determined as a second peer person from the following manners:
and determining the number of the interval persons between the person to be identified and the first co-workers in the first image set or the second image set, and determining the second co-workers in the third image set according to the number of the interval persons. Specifically, the number of the persons to be identified and the number of the persons in the interval between the first person in the same row is recorded as L1, and in the third image set, other persons with the number of the persons in the interval between the persons in the same row and the first person in the same row equal to L1 are searched and preferentially taken as the second person in the same row; other people who have a separation from the first peer equal to L1-1, L1+1, L1-2, L1+2 may also be subsequently determined to be the second peer.
And preferentially determining other people with more simultaneous occurrence times with the first person in the third image set as the second person in the same group. Specifically, the measurement may be performed by the simultaneous occurrence ratio, for example, if some other person C appears in N1 images of the third image set together with the first person B in the third image set, and the third image set has N0 images, the simultaneous occurrence ratio of B and C is N1/N0. And preferentially determining the persons with high simultaneous occurrence ratio as the second fellow persons.
In practical applications, different manners may be combined, for example, the number of persons spaced from the first peer in the third image set is equal to L1, and the occurrence ratio is greater than 50%, and the other persons whose face image clarity detection result meets the first quality condition and whose human body image clarity detection result meets the second quality condition are the second peer.
In step S240, the image of the person to be identified in the first image set is compared with the image of the second fellow person in the third image set to determine whether the person to be identified is the second fellow person.
In this exemplary embodiment, it may be considered that the peer relationship among different persons in the first image set, the second image set and the third image set does not change significantly, that is, the person to be identified in the first image set and the first peer person in the second image set are in the same row, and both still have the peer relationship in the third image set, so that the person to be identified and the second peer person may be the same person. Generally, the face of the person to be identified in the first image set is blocked or unclear, the face and the human body of the second person in the same row in the third image set are clear, and human body characteristics of the first person and the second person in the same row, such as body shape characteristics, body state characteristics, image characteristics of clothes and the like, can be compared to determine whether the persons are the same person.
In an alternative embodiment, step S240 may include the following steps:
detecting the similarity between the human body image of the person to be identified and the human body image of the second person in the same row; for example, the human body image of the person to be identified may be captured from the first image set, the human body image of the second person in the same row may be captured from the third image set, and the body shape feature, the body state feature, the image feature of the clothes, and the like may be extracted from the two human body images to form two feature vectors, and then the similarity between the two feature vectors may be calculated.
And when the similarity is greater than a fifth similarity threshold, determining that the person to be identified and the second peer person are the same person. The fifth similarity threshold is a similarity requirement to be met by determining that two human body images are the same person, and may be set according to experience or actual requirements, for example, 80%, and when the similarity is greater than 80%, it is determined that the person to be identified is a second peer person. The identity of the second peer can be further determined by face recognition of the second peer. For example, the face image of the second fellow person in the third image set is compared with the face image in the face image library to determine the identity of the second fellow person, that is, the identity of the person to be identified.
In an alternative embodiment, a plurality of first persons in common may be determined from the second image set, for example, first person in common D, E, F may be determined, and for each first person in common, at least one corresponding second person in common may be determined from the third image set, for example, D1, D2, D3 for D, E1, E2 for E, and F1, F2, F3, F4 for F. Based on this, referring to fig. 3, step S240 may be implemented by the following steps S310 to S330:
step S310, determining the image of the second peer with the highest image similarity with the person to be identified as the image of the candidate person in the second peers corresponding to each first peer respectively. Namely, determining one with the highest similarity to the person A to be identified in second peer people D1, D2 and D3 corresponding to the first peer people D, and assuming the similarity to be D2; determining one of second peer personnel E1 and E2 corresponding to E with the highest similarity to A, and assuming the similarity to A to be E1; and determining one of second persons in the same group F1, F2, F3 and F4 corresponding to the F with the highest similarity to the A, and assuming that the similarity is F4. D2, E1 and F4 are candidate persons.
Step S320, clustering the images of the candidate persons so that the candidate persons in each category are the same person. Specifically, the similarity may be calculated for each two images of the candidate, and the candidate in the two images with higher similarity (for example, greater than a fifth similarity threshold preset according to experience or actual requirements) is determined to be the same person and classified into the same category. For example, D2 has a high similarity to E1, D2 has a low similarity to F4, and E1 has a low similarity to F4, then D2 and E1 consider the same person and classify into one category, and F4 classifies into another category.
And step S330, determining that the person to be identified and the candidate person in the maximum category are the same person. The maximum category refers to a category having the largest number of images including the number of candidates, and for example, if D2 and E1 are the maximum category, it is determined that the person a to be recognized is the same person as D2 and E1.
In the manner of fig. 3, the accuracy of the person identification is further improved.
In practical application, a second person in the same row as the person to be identified can be searched in a mode of determining the persons in the same row one by one and comparing the persons in the same row. Specifically, reference may be made to the flow in fig. 4:
step S401, a first image set is obtained, wherein the first image set comprises at least one image of a person to be identified.
Step S402, a second reference image is obtained according to a first preset condition, and a second image set is obtained.
In step S403, a first peer is determined from the second image set, for example, other people with a small number of people apart from the people to be identified, a large number of times of appearance in the second image set, and a sharpness detection result of the face image meeting the first quality condition may be preferentially selected as the first peer.
And S404, acquiring a third reference image according to a third preset condition to obtain a third image set.
In step S405, a second peer is determined from the third image set, for example, the number of persons spaced from the first peer is equal to or similar to the number of persons spaced from the first peer and the person to be identified, the number of times of persons co-present with the first peer in the third image set is large, the clear detection result of the human face image satisfies the first quality condition, and the other persons whose clear detection result of the human body image satisfies the second quality condition may be preferentially selected as the second peer.
Step S406, calculating the similarity between the human body image of the person to be identified and the human body image of the second person in the same row.
Step S407, determining whether the person to be identified and the second peer person are the same person according to the similarity, if so, executing step S408; if not, step S409 is executed.
Step S408, determining that the person to be identified and the second peer person are the same person, thereby determining the identity of the person to be identified, and ending the process.
Step S409, detecting whether the second fellow staff in the third image set is exhausted, if so, executing step S410; if not, the process goes to step S405, and another second peer is replaced in the third image set.
Step S410, detecting whether the first peer in the second image set is exhausted, if yes, jumping to step S402, and replacing the first peer with another second image set; if not, the process goes to step S403, and the second image set is replaced with another first peer.
Through the process, the first peer in the second image set and the second peer in the third image set can be traversed, one second peer is selected each time to be compared with the to-be-identified peer in the human body image, when the identity is determined, the identity is successfully identified, and the process is ended. This can reduce the amount of calculation and improve the processing efficiency.
Exemplary embodiments of the present disclosure also provide another person identification method. Referring to fig. 5, the method includes the following steps S510 to S540:
step S510, a first image set including a person to be recognized is obtained, where the image of the person to be recognized includes a face of the person to be recognized that does not satisfy a first quality condition or does not include a face of the person to be recognized, and a human body including the person to be recognized that satisfies a second quality condition.
This step is substantially the same as the implementation of step S210, and thus is not described in detail.
Step S520, determining a second image set containing images of at least one reference person according to the first image set; the image of at least one reference person in the second image set comprises the face of the reference person meeting the first quality condition and the human body of the reference person meeting the second quality condition, and the human body of the reference person in the image of at least one reference person is similar to the human body of the person to be recognized in the image of the person to be recognized.
This step is substantially the same as the implementation process of step S220, and for example, a second reference image may be determined according to a first preset condition, and then other images satisfying the second preset condition with the second reference image may be acquired to form a second image set. The difference lies in that: in step S520, when the second image set is determined, the human body image of the reference person and the human body image of the person to be identified need to be detected, the second image set needs to include at least one image, and the similarity between the human body image and the human body image of the person to be identified reaches a sixth similarity threshold, for example, the body shape feature, the body state feature, the image feature of the clothes, and the like can be respectively extracted to form two feature vectors, and the similarity is calculated to determine whether the similarity reaches the sixth similarity threshold. The sixth similarity threshold is another similarity requirement to be met by determining that the two human body images are the same person, and may be the same as or different from the fifth similarity threshold; the sixth similarity threshold may be determined empirically or by actual requirements, and may be 70% for example. And if the sixth similarity threshold is reached, determining the corresponding person in the second image set as the reference person. The reference person may be any one or more persons similar to the person to be identified in the second image set, for example, a person with the highest similarity to the person to be identified in the second image set may be selected.
For example, if the first image set is obtained by shooting by the monitoring camera P1 in the time period from t0-10s to t0+10s, the monitoring camera P2 closest to P1 can be searched, the images shot by the monitoring camera P2 in the time period from t0+30s to t0+50s are obtained, a second image set is formed, and whether people similar to the people to be identified exist or not is searched; if the information is found, determining the information as a reference person; if the person is not found, the time period from t0+30s to t0+50s is moved to another time period, and the image shot by the P2 is obtained until the person similar to the person to be identified is found.
Step S530, at least one image of a first peer and at least one image of a second peer are obtained according to the first image set and/or the second image set, the first peer is the peer of the person to be identified, and the second peer is the peer of the reference person.
For the determination method of the first peer and the second peer, reference may be made to the contents of the above steps S220 and S230, and thus, the details are not described again.
In an alternative embodiment, for example, a plurality of reference persons are determined, and for each reference person, a corresponding second peer person may be determined. For example, 4 reference persons D, E, F, G similar to the person a to be identified are determined, second peer persons D1 and D2 corresponding to D can be determined, second peer person E1 corresponding to E can be determined, second peer persons F1, F2 and F3 corresponding to F can be determined, and second peer persons G1 and G2 corresponding to G can be determined.
Furthermore, the number of the second peer corresponding to each reference person can be the same as that of the first peer. For example, 3 first peer persons a1, a2, A3 are determined for the person a to be identified, for the reference person D, E, F, G, 3 second peer persons are determined for each reference person, respectively, for example, the second peer persons D1, D2, D3 corresponding to D are determined, the second peer persons E1, E2, E3 corresponding to E are determined, the second peer persons F1, F2, F3 corresponding to F are determined, and the second peer persons G1, G2, G3 corresponding to G are determined.
And step S540, comparing the image of the first peer with the image of the second peer to determine whether the person to be identified is a reference person.
Generally, when the image similarity between the image of the first peer and the image of the second peer is higher, for example, greater than a seventh similarity threshold, it is determined that the first peer and the second peer are the same person, and it is further determined that the person to be identified and the reference person are the same person. The seventh similarity threshold is a similarity requirement which is determined to be met by the same person by the persons in the two images, and may be the same as or different from the fifth similarity threshold and the sixth similarity threshold; the seventh similarity threshold may be determined empirically or in real demand, and in general, the seventh similarity threshold is for an image containing a human face, and therefore a higher value may be set, such as 85%. Under the condition that the person to be identified and the reference person are determined to be the same person, the person to be identified can be further identified through the identity of the reference person.
In an alternative embodiment, the first peer may be compared with the second peer corresponding to each reference person, for example, the first peer { a1, a2, A3} is compared with { D1, D2, D3}, { E1, E2, E3}, { F1, F2, F3}, { G1, G2, G3} respectively, so that the similarity between { a1, a2, A3} and { E1, E2, E3} is the highest, and it is determined that the person a to be identified and the reference person E are the same person.
By the method of fig. 5, the association between the person to be recognized and the reference person can be established according to the relation of the same row of persons under the condition that the face of the person to be recognized is blocked or unclear, and the identity of the person to be recognized can be further determined. On the one hand, the problem that the identity of a person cannot be determined when the face is shielded or unclear in the related technology is solved, and the accuracy of person identification is improved. On the other hand, extra information is not needed, the implementation process is simple, and the practicability is high.
Exemplary embodiments of the present disclosure also provide a person identification apparatus. As shown in fig. 6, the person identifying apparatus 600 may include:
the image acquisition module 610 is configured to acquire a first image set including images of a person to be recognized, where the images of the person to be recognized include faces of the person to be recognized that do not satisfy a first quality condition or faces of the person not to be recognized, and include a human body of the person to be recognized that satisfy a second quality condition;
a first determining module 620, configured to determine a second image set including images of at least one first peer according to the first image set, where the first peer is a peer of the person to be identified;
a second determining module 630, configured to determine a third image set including images of at least one second fellow person from the second image set, the second fellow person being a fellow person of the first fellow person; the image of at least one second peer in the third image set comprises the face of the second peer meeting the first quality condition and the human body of the second peer meeting the second quality condition;
and the person comparison module 640 is configured to compare the image of the person to be identified in the first image set with the image of the second peer person in the third image set to determine whether the person to be identified is the second peer person.
In an alternative embodiment, the first determining module 620 is configured to:
acquiring a second reference image which meets a first preset condition with the images in the first image set; the second image set comprises a second reference image;
the first preset condition includes any one or more of:
the difference of the shooting time is smaller than a first time threshold;
the distance between the shooting places is smaller than a first distance threshold value;
the scene similarity is greater than a first similarity threshold.
In an alternative embodiment, the first determining module 620 is configured to:
acquiring other images meeting a second preset condition with the second reference image, wherein the second image set also comprises the other images meeting the second preset condition;
the second preset condition includes any one or more of:
the second reference image and other images meeting a second preset condition belong to a personnel image set of the same person, and the personnel image set is divided in advance;
the similarity between the face image of the first person in the second reference image and the face images in the other images meeting the second preset condition is greater than a second similarity threshold.
In an alternative embodiment, the second determining module 630 is configured to:
acquiring a third reference image meeting a third preset condition with the images in the second image set, wherein the third image set comprises a third reference image;
the third preset condition includes any one or more of:
the difference of the shooting time is smaller than a third time threshold;
the distance between the shooting places is smaller than a third distance threshold;
the scene similarity is greater than a third similarity threshold.
In an alternative embodiment, the second determining module 630 is configured to:
acquiring other images meeting a fourth preset condition with the third reference image, wherein the third image set also comprises the other images meeting the fourth preset condition;
the fourth preset condition includes any one or more of:
the third reference image and other images meeting a third preset condition belong to a personnel image set of the same person, and the personnel image set is divided in advance;
the similarity between the face image of the second person in the third reference image and the face images in other images under the third preset condition is greater than a fourth similarity threshold.
In an alternative embodiment, the first determination module 620 is configured to determine the at least one first peer from the second image set by:
and in the second image set, carrying out clear detection on the face images of people except the people to be recognized, and determining at least one person of which the clear detection result of the face image meets a first quality condition as a first peer.
In an alternative embodiment, the second determination module 630 is configured to determine at least one second peer person from the third image set by:
and in the third image set, carrying out clear detection on the face images and the human body images of people except the first peer, and determining at least one person of which the clear detection result of the face image meets the first quality condition and the clear detection result of the human body image meets the second quality condition as a second peer.
In an alternative embodiment, the people comparison module 640 is configured to:
and when the similarity between the human body image of the person to be identified and the human body image of the second peer is detected to be greater than a fifth similarity threshold, determining that the person to be identified and the second peer are the same person.
In an alternative embodiment, when the first determining module 620 determines a plurality of first peer people from the first image set, and the second determining module 630 determines at least one corresponding second peer person for each first peer person from the third image set, the people comparing module 640 is configured to:
determining the image of a second peer with the highest image similarity with the person to be identified as the image of the candidate person in second peers corresponding to each first peer respectively;
clustering images of the candidate persons to enable the candidate persons in each category to be the same person;
and determining that the person to be identified is the same as the candidate in the maximum category.
Exemplary embodiments of the present disclosure also provide another person identification apparatus. As shown in fig. 7, the person identification apparatus 700 may include:
a first obtaining module 710, configured to obtain a first image set including an image of a person to be recognized, where the image of the person to be recognized includes a face of the person to be recognized that does not satisfy a first quality condition or does not include a face of the person to be recognized, and includes a human body of the person to be recognized that satisfies a second quality condition;
an image set determination module 720 for determining a second image set comprising images of at least one reference person from the first image set; the image of at least one reference person in the second image set comprises the face of the reference person meeting the first quality condition and the human body of the reference person meeting the second quality condition, and the human body of the reference person in the image of at least one reference person is similar to the human body of the person to be recognized in the image of the person to be recognized;
the second obtaining module 730 is configured to obtain an image of at least one first peer and an image of at least one second peer according to the first image set and/or the second image set, where the first peer is a peer of the person to be identified, and the second peer is a peer of the reference person;
and the person comparison module 740 is used for comparing the image of the first peer with the image of the second peer to determine whether the person to be identified is the reference person.
The specific details of each part of the above-mentioned person identification apparatus 600 and person identification apparatus 700 have been described in detail in the method part embodiment, and the details that are not disclosed can be referred to the method part embodiment, and thus are not described again.
Exemplary embodiments of the present disclosure also provide a computer-readable storage medium, which may be implemented in the form of a program product, including program code for causing an electronic device to perform the steps according to various exemplary embodiments of the present disclosure described in the above-mentioned "exemplary method" section of this specification, when the program product is run on the electronic device. The program product may employ a portable compact disc read only memory (CD-ROM) and include program code, and may be run on an electronic device, such as a personal computer. However, the program product of the present disclosure is not limited thereto, and in this document, a readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
A computer readable signal medium may include a propagated data signal with readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A readable signal medium may also be any readable medium that is not a readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Program code for carrying out operations for the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., through the internet using an internet service provider).
The exemplary embodiment of the present disclosure also provides an electronic device capable of implementing the above method. An electronic device 800 according to such an exemplary embodiment of the present disclosure is described below with reference to fig. 8. The electronic device 800 shown in fig. 8 is only an example and should not bring any limitations to the functionality and scope of use of the embodiments of the present disclosure.
As shown in fig. 8, electronic device 800 may take the form of a general purpose computing device. The components of the electronic device 800 may include, but are not limited to: at least one processing unit 810, at least one memory unit 820, a bus 830 connecting the various system components including the memory unit 820 and the processing unit 810, and a display unit 840.
The storage unit 820 stores program code that may be executed by the processing unit 810 to cause the processing unit 810 to perform steps according to various exemplary embodiments of the present disclosure described in the "exemplary methods" section above in this specification. For example, processing unit 810 may perform any one or more of the method steps of fig. 2-5.
The storage unit 820 may include readable media in the form of volatile storage units, such as a random access storage unit (RAM)821 and/or a cache storage unit 822, and may further include a read only storage unit (ROM) 823.
The electronic device 800 may also communicate with one or more external devices 900 (e.g., keyboard, pointing device, bluetooth device, etc.), with one or more devices that enable a user to interact with the electronic device 800, and/or with any devices (e.g., router, modem, etc.) that enable the electronic device 800 to communicate with one or more other computing devices. Such communication may occur via input/output (I/O) interfaces 850. Also, the electronic device 800 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the internet) via the network adapter 860. As shown, the network adapter 860 communicates with the other modules of the electronic device 800 via the bus 830. It should be appreciated that although not shown, other hardware and/or software modules may be used in conjunction with the electronic device 800, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.
Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein may be implemented by software, or by software in combination with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a usb disk, a removable hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which may be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the exemplary embodiments of the present disclosure.
Furthermore, the above-described figures are merely schematic illustrations of processes included in methods according to exemplary embodiments of the present disclosure, and are not intended to be limiting. It will be readily understood that the processes shown in the above figures are not intended to indicate or limit the chronological order of the processes. In addition, it is also readily understood that these processes may be performed synchronously or asynchronously, e.g., in multiple modules.
As will be appreciated by one skilled in the art, aspects of the present disclosure may be embodied as a system, method or program product. Accordingly, various aspects of the present disclosure may be embodied in the form of: an entirely hardware embodiment, an entirely software embodiment (including firmware, microcode, etc.) or an embodiment combining hardware and software aspects that may all generally be referred to herein as a "circuit," module "or" system.
It should be noted that although in the above detailed description several modules or units of the device for action execution are mentioned, such a division is not mandatory. Indeed, the features and functions of two or more modules or units described above may be embodied in one module or unit, according to exemplary embodiments of the present disclosure. Conversely, the features and functions of one module or unit described above may be further divided into embodiments by a plurality of modules or units.
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
It will be understood that the present disclosure is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is to be limited only by the following claims.
Claims (14)
1. A person identification method, comprising:
acquiring a first image set containing images of a person to be recognized, wherein the images of the person to be recognized contain faces of the person to be recognized which do not meet a first quality condition or faces of the person to be recognized, and contain bodies of the person to be recognized which meet a second quality condition;
determining a second image set comprising images of at least one first co-worker according to the first image set, wherein the first co-worker is a co-worker of the person to be identified;
determining a third image set comprising images of at least one second peer person from the second image set, the second peer person being a peer person of the first peer person; wherein at least one image of the second peer in the third image set includes a face of the second peer satisfying the first quality condition and a body of the second peer satisfying the second quality condition;
and comparing the image of the person to be identified in the first image set with the image of the second person in the same row in the third image set to determine whether the person to be identified is the second person in the same row.
2. The method of claim 1, wherein determining a second image set comprising images of at least a first peer from the first image set comprises:
acquiring a second reference image meeting a first preset condition with images in the first image set, wherein the second image set comprises the second reference image;
the first preset condition comprises any one or more of the following conditions:
the difference of the shooting time is smaller than a first time threshold;
the distance between the shooting places is smaller than a first distance threshold value;
the scene similarity is greater than a first similarity threshold.
3. The method of claim 2, wherein determining a second image set comprising images of at least a first peer from the first image set further comprises:
acquiring other images meeting a second preset condition with the second reference image, wherein the second image set also comprises the other images meeting the second preset condition;
the second preset condition comprises any one or more of the following conditions:
the second reference image and the other images meeting the second preset condition belong to a personnel image set of the same person, and the personnel image set is divided in advance;
the similarity between the face image of the first person in the second reference image and the face images in the other images meeting the second preset condition is greater than a second similarity threshold.
4. The method of any of claims 1-3, wherein determining a third image set comprising images of at least one second peer from the second image set comprises:
acquiring a third reference image meeting a third preset condition with the images in the second image set, wherein the third image set comprises the third reference image;
the third preset condition includes any one or more of:
the difference of the shooting time is smaller than a third time threshold;
the distance between the shooting places is smaller than a third distance threshold;
the scene similarity is greater than a third similarity threshold.
5. The method of claim 4, wherein determining a third image set comprising images of at least a second peer from the second image set further comprises:
acquiring other images meeting a fourth preset condition with the third reference image, wherein the third image set further comprises the other images meeting the fourth preset condition;
the fourth preset condition includes any one or more of:
the third reference image and the other images meeting the third preset condition belong to a personnel image set of the same person, and the personnel image set is divided in advance;
the similarity between the face image of the second person in the third reference image and the face images in the other images under the third preset condition is greater than a fourth similarity threshold.
6. The method according to any of claims 1-5, characterized in that at least one first peer is determined from the second image set by:
and in the second image set, carrying out clear detection on the face images of people except the people to be identified, and determining at least one person of which the clear detection result of the face image meets the first quality condition as the first person in parallel.
7. Method according to any of claims 1-6, characterized in that at least one second peer person is determined from the third image set by:
and in the third image set, performing clear detection on the face images and the human body images of people except the first peer, and determining at least one person whose face image clear detection result meets the first quality condition and whose human body image clear detection result meets the second quality condition as the second peer.
8. The method according to any one of claims 1-7, wherein said comparing the image of the person to be identified in the first image set with the image of the second peer in the third image set to determine whether the person to be identified is the second peer comprises:
and when the similarity between the human body image of the person to be identified and the human body image of the second peer is detected to be greater than a fifth similarity threshold, determining that the person to be identified and the second peer are the same person.
9. The method according to any one of claims 1 to 8, wherein when a plurality of first persons in a same row are determined from the second image set and a corresponding at least one second person in a same row is determined for each of the first persons in the third image set, the comparing the image of the person to be identified in the first image set with the image of the second person in the third image set to determine whether the person to be identified is the second person comprises:
determining the image of a second peer with the highest image similarity with the person to be identified as the image of the candidate person in second peers corresponding to each first peer respectively;
clustering the images of the candidate persons to enable the candidate persons in each category to be the same person;
and determining that the person to be identified and the candidate person in the category containing the largest number of images are the same person.
10. A person identification method, comprising:
acquiring a first image set containing images of a person to be recognized, wherein the images of the person to be recognized contain faces of the person to be recognized which do not meet a first quality condition or faces of the person to be recognized, and contain bodies of the person to be recognized which meet a second quality condition;
determining a second image set comprising images of at least one reference person from the first image set; wherein at least one image of the reference person in the second image set includes a human face of the reference person satisfying the first quality condition and a human body of the reference person satisfying the second quality condition, and the human body of the reference person in the at least one image of the reference person is similar to the human body of the person to be recognized in the image of the person to be recognized;
acquiring at least one image of a first peer and at least one image of a second peer according to the first image set and/or the second image set, wherein the first peer is the peer of the person to be identified, and the second peer is the peer of the reference person;
and comparing the image of the first peer with the image of the second peer to determine whether the person to be identified is the reference person.
11. A person identification device, comprising:
the image acquisition module is used for acquiring a first image set containing images of people to be recognized, wherein the images of the people to be recognized contain faces of the people to be recognized which do not meet a first quality condition or faces of the people to be recognized, and contain human bodies of the people to be recognized which meet a second quality condition;
a first determining module, configured to determine, according to the first image set, a second image set including images of at least one first peer, where the first peer is a peer of the to-be-identified person;
a second determining module, configured to determine, according to the second image set, a third image set including images of at least one second peer person, where the second peer person is a peer person of the first peer person; wherein at least one image of the second peer in the third image set includes a face of the second peer satisfying the first quality condition and a body of the second peer satisfying the second quality condition;
and the person comparison module is used for comparing the image of the person to be identified in the first image set with the image of the second person in the same row in the third image set so as to determine whether the person to be identified is the second person in the same row.
12. A person identification device, comprising:
the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring a first image set containing images of people to be recognized, and the images of the people to be recognized contain faces of the people to be recognized which do not meet a first quality condition or faces of the people to be recognized and contain bodies of the people to be recognized which meet a second quality condition;
an image set determination module for determining a second image set comprising images of at least one reference person from the first image set; wherein at least one image of the reference person in the second image set includes a human face of the reference person satisfying the first quality condition and a human body of the reference person satisfying the second quality condition, and the human body of the reference person in the at least one image of the reference person is similar to the human body of the person to be recognized in the image of the person to be recognized;
a second obtaining module, configured to obtain an image of at least one first peer and an image of at least one second peer according to the first image set and/or the second image set, where the first peer is a peer of the to-be-identified person, and the second peer is a peer of the reference person;
and the personnel comparison module is used for comparing the image of the first peer with the image of the second peer so as to determine whether the personnel to be identified is the reference personnel.
13. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the method of any one of claims 1 to 10.
14. An electronic device, comprising:
a processor; and
a memory for storing executable instructions of the processor;
wherein the processor is configured to perform the method of any of claims 1 to 10 via execution of the executable instructions.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010923634.XA CN112232113B (en) | 2020-09-04 | 2020-09-04 | Person identification method, person identification device, storage medium, and electronic apparatus |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010923634.XA CN112232113B (en) | 2020-09-04 | 2020-09-04 | Person identification method, person identification device, storage medium, and electronic apparatus |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112232113A true CN112232113A (en) | 2021-01-15 |
CN112232113B CN112232113B (en) | 2022-08-05 |
Family
ID=74116998
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010923634.XA Active CN112232113B (en) | 2020-09-04 | 2020-09-04 | Person identification method, person identification device, storage medium, and electronic apparatus |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112232113B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113553965A (en) * | 2021-07-27 | 2021-10-26 | 重庆电子工程职业学院 | Person identity recognition method combining face recognition and human body recognition |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150169946A1 (en) * | 2013-12-12 | 2015-06-18 | Evernote Corporation | User discovery via digital id and face recognition |
CN109117714A (en) * | 2018-06-27 | 2019-01-01 | 北京旷视科技有限公司 | A kind of colleague's personal identification method, apparatus, system and computer storage medium |
CN110084103A (en) * | 2019-03-15 | 2019-08-02 | 深圳英飞拓科技股份有限公司 | A kind of same pedestrian's analysis method and system based on face recognition technology |
CN110276272A (en) * | 2019-05-30 | 2019-09-24 | 罗普特科技集团股份有限公司 | Confirm method, apparatus, the storage medium of same administrative staff's relationship of label personnel |
WO2019223313A1 (en) * | 2018-05-22 | 2019-11-28 | 深圳云天励飞技术有限公司 | Personnel file establishment method and apparatus |
-
2020
- 2020-09-04 CN CN202010923634.XA patent/CN112232113B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150169946A1 (en) * | 2013-12-12 | 2015-06-18 | Evernote Corporation | User discovery via digital id and face recognition |
WO2019223313A1 (en) * | 2018-05-22 | 2019-11-28 | 深圳云天励飞技术有限公司 | Personnel file establishment method and apparatus |
CN109117714A (en) * | 2018-06-27 | 2019-01-01 | 北京旷视科技有限公司 | A kind of colleague's personal identification method, apparatus, system and computer storage medium |
CN110084103A (en) * | 2019-03-15 | 2019-08-02 | 深圳英飞拓科技股份有限公司 | A kind of same pedestrian's analysis method and system based on face recognition technology |
CN110276272A (en) * | 2019-05-30 | 2019-09-24 | 罗普特科技集团股份有限公司 | Confirm method, apparatus, the storage medium of same administrative staff's relationship of label personnel |
Non-Patent Citations (1)
Title |
---|
陈建旭等: "基于视频的人像识别技战法应用", 《中国公共安全(学术版)》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113553965A (en) * | 2021-07-27 | 2021-10-26 | 重庆电子工程职业学院 | Person identity recognition method combining face recognition and human body recognition |
CN113553965B (en) * | 2021-07-27 | 2023-09-12 | 重庆电子工程职业学院 | Person identity recognition method combining face recognition and human body recognition |
Also Published As
Publication number | Publication date |
---|---|
CN112232113B (en) | 2022-08-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107292240B (en) | Person finding method and system based on face and body recognition | |
CN110175549B (en) | Face image processing method, device, equipment and storage medium | |
CN108921008B (en) | Portrait identification method and device and electronic equipment | |
KR101687530B1 (en) | Control method in image capture system, control apparatus and a computer-readable storage medium | |
US10795928B2 (en) | Image search apparatus, system, and method | |
CN111898486B (en) | Monitoring picture abnormality detection method, device and storage medium | |
CN111401205A (en) | Action recognition method and device, electronic equipment and computer readable storage medium | |
CN110059666B (en) | Attention detection method and device | |
EP3139308A1 (en) | People search system and people search method | |
CN111723773B (en) | Method and device for detecting carryover, electronic equipment and readable storage medium | |
CN110889314B (en) | Image processing method, device, electronic equipment, server and system | |
CN111241872A (en) | Video image shielding method and device | |
CN114005140A (en) | Personnel identification method, device, equipment, pedestrian monitoring system and storage medium | |
CN112232113B (en) | Person identification method, person identification device, storage medium, and electronic apparatus | |
CN113920585A (en) | Behavior recognition method and device, equipment and storage medium | |
CN114022531A (en) | Image processing method, electronic device, and storage medium | |
US9286707B1 (en) | Removing transient objects to synthesize an unobstructed image | |
US11074696B2 (en) | Image processing device, image processing method, and recording medium storing program | |
JP2012212238A (en) | Article detection device and stationary-person detection device | |
CN111708907A (en) | Target person query method, device, equipment and storage medium | |
JPWO2018179119A1 (en) | Video analysis device, video analysis method, and program | |
CN113947795B (en) | Mask wearing detection method, device, equipment and storage medium | |
CN114882073A (en) | Target tracking method and apparatus, medium, and computer device | |
CN115115976A (en) | Video processing method and device, electronic equipment and storage medium | |
CN114387670A (en) | Gait recognition method and device based on space-time feature fusion and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |