CN112419249B - Special clothing picture conversion method, terminal device and storage medium - Google Patents
Special clothing picture conversion method, terminal device and storage medium Download PDFInfo
- Publication number
- CN112419249B CN112419249B CN202011263797.6A CN202011263797A CN112419249B CN 112419249 B CN112419249 B CN 112419249B CN 202011263797 A CN202011263797 A CN 202011263797A CN 112419249 B CN112419249 B CN 112419249B
- Authority
- CN
- China
- Prior art keywords
- picture
- special
- self
- pictures
- coding network
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000006243 chemical reaction Methods 0.000 title claims abstract description 29
- 230000011218 segmentation Effects 0.000 claims abstract description 29
- 238000005286 illumination Methods 0.000 claims abstract description 10
- 230000008859 change Effects 0.000 claims abstract description 9
- 238000009499 grossing Methods 0.000 claims abstract description 9
- 230000009466 transformation Effects 0.000 claims abstract description 7
- 238000004590 computer program Methods 0.000 claims description 18
- 230000006870 function Effects 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 210000000746 body region Anatomy 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 3
- 230000001131 transforming effect Effects 0.000 claims description 2
- 230000004807 localization Effects 0.000 claims 1
- 238000002372 labelling Methods 0.000 description 4
- 238000011176 pooling Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 230000036544 posture Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000011426 transformation method Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- 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
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/30108—Industrial image inspection
- G06T2207/30124—Fabrics; Textile; Paper
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention relates to a special clothing picture conversion method, terminal equipment and a storage medium, wherein the method comprises the following steps: collecting pictures containing special clothes to form a training set; carrying out example segmentation, Gaussian smoothing, Hadamard product operation and color change on the pictures in the training set, and then carrying out image superposition operation on the pictures after color change and the original pictures; taking all the superposed pictures as the input of a self-coding network, and training the self-coding network to ensure that the difference of corresponding original pictures in a picture training set output by the self-coding network is minimum; and after the picture of the special clothes to be identified is converted through the trained self-coding network, identifying the special clothes. The method avoids the influence of the color transformation such as illumination, contrast or tone of the picture on the identification of the special clothes in the picture by training the self-coding network, and improves the overall identification rate of the identification of the special clothes.
Description
Technical Field
The invention relates to the field of image processing, in particular to a special clothing picture conversion method, terminal equipment and a storage medium.
Background
Although the special clothing identification method using the convolutional neural network has been significantly improved, the network model is trained by using pictures of simple scenes downloaded from the internet, and the special clothing identification precision is challenged along with the changes of illumination, visual angles and clothing shapes in complex scenes.
Disclosure of Invention
In order to solve the above problems, the present invention provides a special clothing image transformation method, a terminal device and a storage medium.
The specific scheme is as follows:
a special clothes picture conversion method comprises the following steps:
s1: collecting pictures containing special clothes to form a training set;
s2: carrying out example segmentation on the special clothing region of each picture in the training set, and extracting the special clothing region picture after the example segmentation;
s3: performing Gaussian smoothing processing on the special clothing region picture after the example segmentation;
s4: carrying out Hadamard product operation on each picture subjected to Gaussian smoothing and the corresponding original picture in the training set;
s5: carrying out color change on the picture subjected to the Hadamard product operation, and carrying out image superposition operation on the picture subjected to the color change and the corresponding original picture in the training set;
s6: taking all the superposed pictures as the input of a self-coding network, and training the self-coding network to ensure that the difference of corresponding original pictures in a picture training set output by the self-coding network is minimum;
s7: and after the picture of the special clothes to be identified is converted through the trained self-coding network, identifying the special clothes.
Further, the example segmentation specifically comprises the following steps:
s21: carrying out weak positioning on the special clothing area on the picture;
s22: performing semantic segmentation on a human body region in the image through an image semantic segmentation algorithm;
s23: and performing intersection and comparison IOU calculation on the weak positioning area and the semantic segmentation area, and judging the semantic segmentation area as a special clothing area picture to be extracted when the IOU is greater than a threshold value.
Further, weak positioning is carried out by adopting a CAM technology of generating a class activation map.
Further, the IOU calculation formula is as follows:
wherein u is 1 Is a weakly localized area, u 2 The regions are semantically segmented.
Further, color transformation includes transforming illumination, contrast, or hue.
Further, the specific formula of the image superposition operation is as follows:
I add ＝I c *I mask +I s (1-I mask )
wherein, I add For superimposed pictures, I c For pictures after color conversion, I mask Splitting the picture for example I s Corresponding original pictures in the training set.
Further, the loss function of the self-coding network is L 2 A loss function.
The special clothing image conversion terminal device comprises a processor, a memory and a computer program which is stored in the memory and can run on the processor, wherein the processor executes the computer program to realize the steps of the method of the embodiment of the invention.
A computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, carries out the steps of the method as described above for an embodiment of the invention.
By adopting the technical scheme, the method avoids the influence of color transformation such as illumination, contrast or tone of the picture on the identification of the special clothes in the picture by training the self-coding network, and improves the overall identification rate of the identification of the special clothes.
Drawings
Fig. 1 is a flowchart illustrating a first embodiment of the present invention.
Fig. 2 is a schematic diagram of the network structure of the self-coding network in this embodiment.
Detailed Description
To further illustrate the various embodiments, the invention provides the accompanying drawings. The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the embodiments. Those skilled in the art will appreciate still other possible embodiments and advantages of the present invention with reference to these figures.
The invention will now be further described with reference to the accompanying drawings and detailed description.
The first embodiment is as follows:
the embodiment of the invention provides a special clothes picture conversion method, as shown in figure 1, the method comprises the following steps:
s1: and collecting pictures containing special clothes to form a training set.
In this embodiment, 10000 special clothes pictures with different scenes and different categories are collected from the internet, and the special clothes are provided with the following categories: military and police clothes are provided with army, air force, navy, rocket force clothes and public security clothes, religious clothes, government specific clothes and the like.
S2: and carrying out example segmentation on the special clothing region of each picture in the training set, and extracting the special clothing region picture after the example segmentation.
Because the pixel-level labeling of the example segmentation is a great workload, although the garment boundary can be labeled manually, the region surrounded by the boundary can be regarded as the region where the garment is located, but the human body has different postures, and the labeling of the trunk and the limbs takes a lot of time. In order to alleviate the problem of manual labeling, the pixel-level semantic labeling problem can be quickly solved by combining a target weak positioning method on the basis of a human body semantic segmentation algorithm.
In this embodiment, the instance segmentation specifically includes the following steps:
s21: and carrying out weak positioning on the special clothing area on the picture.
The weak positioning is to position the special clothing in the image by the label on the image layer, and different from the target detection, the target detection network needs to label the position information of the special clothing, and the weak positioning method only needs to label the type of the image.
In the embodiment, the weak positioning is carried out by adopting a CAM technology of generating class activation graphs. The CAM technology is mostly convolutional layer, and only uses the average pooling layer before the output layer (softmax for classification), and uses the output of the average pooling layer as the input feature of the fully connected layer for classification. With this simple connection structure, important areas in the picture can be marked in a way that the output layer weights are mapped back to the convolutional layer features. The global average pooling layer outputs an average of the feature maps of each cell of the last convolutional layer, and a weighted sum of these values is used to generate the final output. It can also be said that the CAM is obtained by calculating the weighted sum of the last convolutional layer feature map, which highlights the clothing saliency region.
M is used for the feature map after the global average pooling 1 ,M 2 ,…M n And if so, calculating the layer behind the CAM by the following formula:
CAM＝w 1 *M 1 +w 2 *M 2 +…+w n *M n
s22: and performing semantic segmentation on the human body region in the image through an image semantic segmentation algorithm.
Since the open-source image instance segmentation algorithm does not use the special clothes as the output category, in the embodiment, the human body wearing the special clothes is preferentially considered as the output category, and the human body region is segmented by using the image semantic segmentation algorithm.
S23: performing intersection and comparison IOU calculation on the weak positioning area and the semantic segmentation area, and judging the semantic segmentation area as a special clothing area picture to be extracted when the IOU is greater than a threshold value; otherwise, the semantic partition is discarded.
The IOU calculation is as follows:
wherein u is 1 Is a weakly localized area, u 2 The regions are semantically segmented.
S3: and performing Gaussian smoothing processing on the special clothing region picture after the example segmentation.
The gaussian smoothing process is used to eliminate the edge effect of the segmented region.
S4: and carrying out Hadamard product operation on each picture after the Gaussian smoothing and the corresponding original picture in the training set.
The background of the picture after the Hadamard product operation is black, and the foreground of the picture is a human body region picture after Gaussian smoothing processing.
S5: and carrying out color change on the picture after the Hadamard product operation, and carrying out image superposition operation on the picture after the color change and the corresponding original picture in the training set.
The color transformation can be illumination, contrast or tone transformation, and the like, and is not limited herein, the illumination, contrast or tone of the picture and the like are adjusted through color change, and the self-coding network convenient for subsequent training can convert the picture which cannot be subjected to special clothing identification due to abnormal illumination, contrast or tone into the picture which is normal in illumination, contrast or tone and can be subjected to special clothing identification.
The specific formula of the image superposition operation is as follows:
I add ＝I c *I mask +I s (1-I mask )
wherein, I add For superimposed pictures, I c As a picture after color conversion, I mask Splitting the picture for example, I s Is the corresponding original picture in the training set.
Steps S3-S5 are the migration operation of the divided region picture. Through the steps, the superposed picture is more real and natural, and the details such as texture, edge, image style and the like similar to the original picture are kept. The self-coding network can be trained better.
S6: and taking all the superposed pictures as the input of the self-coding network, and training the self-coding network to ensure that the difference of the corresponding original pictures in the picture training set output by the self-coding network is minimum.
The network structure of the self-coding network is shown in fig. 2, wherein c1 represents a convolution layer, the convolution kernel is 1 × 1, and the step size is 2; c2 is also a convolution layer with a convolution kernel of 4 × 4; dc1 is an deconvolution layer, the convolution kernel is 1 × 1, and the step length is 2; dc2 is also the deconvolution layer, with a convolution kernel of 4 x 4. The network structure adopts a characteristic pyramid mode during encoding and decoding, so that the generated image is closer to the original image.
The loss function of the self-coding network is L 2 A loss function, calculated as:
s7: and after the pictures of the special clothes to be recognized are converted through the trained self-coding network, the special clothes are recognized.
Further, in order to ensure the quality of the picture for identifying the special clothes converted by the self-coding network, the embodiment further comprises the step of performing quality judgment on the picture output by the self-coding network, wherein the quality judgment method is performed by adopting a parameterized model GMM-GIQA.
The parameterized model quality judgment method uses a Gaussian Mixture Model (GMM) to fit the characteristics of real data distribution at a characteristic level. For a picture to be detected, extracting picture characteristics x by using a traditional machine learning algorithm or a convolutional neural network, and then inputting the characteristics into the following formula, wherein the output probability is the quality of the picture.
Where x is a picture feature, g (μ) i ,σ i ) Denotes the ith Gaussian model, μ i Is a mean value, σ i Is the variance, w i The sum of all gaussian model weights is equal to 1, which is the weight of the mixture gaussian model.
According to the embodiment of the invention, the influence of color transformation such as illumination, contrast or tone of the picture on the identification of the special clothes in the picture is avoided by training the self-coding network, and the overall identification rate of the identification of the special clothes is improved. The method reduces the image synthesis cost, accelerates the sample synthesis speed, increases the proportion of special clothes in a complex scene, relieves the problem of long tail distribution of training samples, and improves the accuracy of the special clothes identification network. By using the characteristic pyramid self-coding network and combining a human body clothing mask method, the accuracy of special clothing image generation is improved, and the generated image is natural and real, so that the clothing category of the generated image is changed, and other areas of the image keep the original style.
Example two:
the invention also provides special clothing image conversion terminal equipment, which comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, wherein the processor executes the computer program to realize the steps in the method embodiment of the first embodiment of the invention.
Further, as an executable scheme, the special clothing image conversion terminal device may be a desktop computer, a notebook, a palm computer, a cloud server, or other computing devices. The special clothing image conversion terminal equipment can comprise, but is not limited to, a processor and a memory. It is understood by those skilled in the art that the above-mentioned constituent structure of the special clothing picture conversion terminal device is only an example of the special clothing picture conversion terminal device, and does not constitute a limitation on the special clothing picture conversion terminal device, and may include more or less components than the above, or combine some components, or different components, for example, the special clothing picture conversion terminal device may further include an input/output device, a network access device, a bus, and the like, which is not limited in this embodiment of the present invention.
Further, as an executable solution, the Processor may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, a discrete hardware component, and the like. The general processor can be a microprocessor or the processor can be any conventional processor, etc., the processor is the control center of the special clothing image conversion terminal device, and various interfaces and lines are used for connecting all parts of the whole special clothing image conversion terminal device.
The memory can be used for storing the computer program and/or the module, and the processor realizes various functions of the special clothing image conversion terminal equipment by running or executing the computer program and/or the module stored in the memory and calling the data stored in the memory. The memory can mainly comprise a program storage area and a data storage area, wherein the program storage area can store an operating system and an application program required by at least one function; the storage data area may store data created according to the use of the mobile phone, and the like. In addition, the memory may include high speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other volatile solid state storage device.
The invention also provides a computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, carries out the steps of the above-mentioned method of an embodiment of the invention.
The module/unit integrated with the special clothing image conversion terminal device can be stored in a computer readable storage medium if the module/unit is realized in the form of a software functional unit and sold or used as an independent product. Based on such understanding, all or part of the flow of the method according to the embodiments of the present invention may also be implemented by a computer program, which may be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the method embodiments may be implemented. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), software distribution medium, and the like.
While the invention has been particularly shown and described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (9)
1. A special clothes picture conversion method is characterized by comprising the following steps:
s1: collecting pictures containing special clothes to form a training set;
s2: carrying out example segmentation on the special clothing region of each picture in the training set, and extracting the special clothing region picture after the example segmentation;
s3: carrying out Gaussian smoothing processing on the special clothing region picture after the example segmentation;
s4: carrying out Hadamard product operation on each picture subjected to Gaussian smoothing and the corresponding original picture in the training set;
s5: carrying out color change on the picture after the Hadamard product operation, and carrying out image superposition operation on the picture after the color change and the corresponding original picture in the training set;
s6: taking all the superposed pictures as the input of a self-coding network, and training the self-coding network to ensure that the difference of corresponding original pictures in a picture training set output by the self-coding network is minimum;
s7: and after the pictures of the special clothes to be recognized are converted through the trained self-coding network, the special clothes are recognized.
2. The special clothing image conversion method according to claim 1, characterized in that: the example segmentation specifically comprises the following steps:
s21: carrying out weak positioning on the special clothing area on the picture;
s22: performing semantic segmentation on a human body region in the image through an image semantic segmentation algorithm;
s23: and performing intersection and comparison IOU calculation on the weak positioning area and the semantic segmentation area, and judging the semantic segmentation area as a special clothing area picture to be extracted when the IOU is greater than a threshold value.
3. The special clothing image conversion method according to claim 2, characterized in that: weak localization is performed using a CAM technique that generates class activation maps.
4. The special clothes picture conversion method according to claim 2, characterized in that: the IOU calculation formula is:
wherein u is 1 Is a weakly localized area u 2 The regions are semantically partitioned.
5. The special clothing image conversion method according to claim 1, characterized in that: color transformation includes transforming illumination, contrast, or hue.
6. The special clothes picture conversion method according to claim 1, characterized in that: the specific formula of the image superposition operation is as follows:
I add ＝I c *I mask +I s (1-I mask )
wherein, I add For superimposed pictures, I c As a picture after color conversion, I mask Splitting the picture for example, I s Is the corresponding original picture in the training set.
7. The special clothing image conversion method according to claim 1, characterized in that: the loss function of the self-coding network is L 2 A loss function.
8. A special dress picture conversion terminal equipment which characterized in that: comprising a processor, a memory and a computer program stored in said memory and running on said processor, said processor implementing the steps of the method according to any one of claims 1 to 7 when executing said computer program.
9. A computer-readable storage medium storing a computer program, the computer program characterized in that: the computer program when executed by a processor implementing the steps of the method as claimed in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011263797.6A CN112419249B (en) | 2020-11-12 | 2020-11-12 | Special clothing picture conversion method, terminal device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011263797.6A CN112419249B (en) | 2020-11-12 | 2020-11-12 | Special clothing picture conversion method, terminal device and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112419249A CN112419249A (en) | 2021-02-26 |
| CN112419249B true CN112419249B (en) | 2022-09-06 |
Family
ID=74832218
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011263797.6A Active CN112419249B (en) | 2020-11-12 | 2020-11-12 | Special clothing picture conversion method, terminal device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112419249B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114445632A (en) * | 2022-02-08 | 2022-05-06 | 支付宝(杭州)信息技术有限公司 | Picture processing method and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521565A (en) * | 2011-11-23 | 2012-06-27 | 浙江晨鹰科技有限公司 | Garment identification method and system for low-resolution video |
| CN110807367A (en) * | 2019-10-05 | 2020-02-18 | 上海淡竹体育科技有限公司 | Method for dynamically identifying personnel number in motion |
| WO2020107687A1 (en) * | 2018-11-27 | 2020-06-04 | 邦鼓思电子科技（上海）有限公司 | Vision-based working area boundary detection system and method, and machine equipment |
| CN111325806A (en) * | 2020-02-18 | 2020-06-23 | 苏州科达科技股份有限公司 | Clothing color recognition method, device and system based on semantic segmentation |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8068676B2 (en) * | 2007-11-07 | 2011-11-29 | Palo Alto Research Center Incorporated | Intelligent fashion exploration based on clothes recognition |
| KR101844371B1 (en) * | 2014-03-19 | 2018-04-02 | 삼성전자주식회사 | Method and apparatus for processing image |
-
2020
- 2020-11-12 CN CN202011263797.6A patent/CN112419249B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521565A (en) * | 2011-11-23 | 2012-06-27 | 浙江晨鹰科技有限公司 | Garment identification method and system for low-resolution video |
| WO2020107687A1 (en) * | 2018-11-27 | 2020-06-04 | 邦鼓思电子科技（上海）有限公司 | Vision-based working area boundary detection system and method, and machine equipment |
| CN110807367A (en) * | 2019-10-05 | 2020-02-18 | 上海淡竹体育科技有限公司 | Method for dynamically identifying personnel number in motion |
| CN111325806A (en) * | 2020-02-18 | 2020-06-23 | 苏州科达科技股份有限公司 | Clothing color recognition method, device and system based on semantic segmentation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112419249A (en) | 2021-02-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11151363B2 (en) | Expression recognition method, apparatus, electronic device, and storage medium | |
| Kamal et al. | Automatic traffic sign detection and recognition using SegU-Net and a modified Tversky loss function with L1-constraint | |
| CN109359538B (en) | Training method of convolutional neural network, gesture recognition method, device and equipment | |
| Radenovic et al. | Deep shape matching | |
| Flores et al. | Application of convolutional neural networks for static hand gestures recognition under different invariant features | |
| Xiao et al. | Deep salient object detection with dense connections and distraction diagnosis | |
| Li et al. | Weighted feature pyramid networks for object detection | |
| CN111814794B (en) | Text detection method and device, electronic equipment and storage medium | |
| Hu et al. | Image classification using multiscale information fusion based on saliency driven nonlinear diffusion filtering | |
| CN113505768A (en) | Model training method, face recognition method, electronic device and storage medium | |
| CN113490947A (en) | Detection model training method and device, detection model using method and storage medium | |
| CN110852327A (en) | Image processing method, image processing device, electronic equipment and storage medium | |
| CN110222572A (en) | Tracking, device, electronic equipment and storage medium | |
| CN111108508A (en) | Facial emotion recognition method, intelligent device and computer-readable storage medium | |
| Mohseni et al. | Recognizing induced emotions with only one feature: A novel color histogram-based system | |
| CN110610131B (en) | Face movement unit detection method and device, electronic equipment and storage medium | |
| CN112419249B (en) | Special clothing picture conversion method, terminal device and storage medium | |
| CN110659631B (en) | License plate recognition method and terminal equipment | |
| CN117252947A (en) | Image processing method, image processing apparatus, computer, storage medium, and program product | |
| CN116912924A (en) | Target image recognition method and device | |
| Kakkar | Facial expression recognition with LDPP & LTP using deep belief network | |
| Zhao et al. | Micro-expression recognition based on pixel residual sum and cropped gaussian pyramid | |
| CN112801045B (en) | Text region detection method, electronic equipment and computer storage medium | |
| CN112785601B (en) | Image segmentation method, system, medium and electronic terminal | |
| CN114283087A (en) | Image denoising method and related equipment |
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 |