Clothes segmentation method and device, electronic equipment and storage medium

Abstract

The invention discloses a clothes segmentation method and apparatus, an electronic device and a storage medium. The method comprises the steps of obtaining a to-be-detected clothes video; performing image screening on the to-be-detected clothes video, and determining a plurality of to-be-segmented clothes images; and synchronously inputting the plurality of to-be-segmented clothes images into a clothes segmentation model, and outputting a clothes segmentation result corresponding to each to-be-segmented clothes image by using the clothes segmentation model. The method can effectively cope with the synchronous segmentation of a video multi-frame scene and a plurality of clothes images, achieves the automatic segmentation of the clothes images, improves the clothes segmentation efficiency and accuracy, improves the clothes segmentation effect, and can be widely applied to the technical field of image processing.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and particularly relates to a clothing segmentation method, device, electronic device, and storage medium. Background Art

[0002] Clothing image segmentation can provide a good carrier for subsequent applications such as clothing fashion understanding and clothing management. In existing clothing segmentation methods, an image segmentation model with multiple branches is usually used to integrate the segmentation results of clothing images, thereby introducing more parameters, resulting in a slower speed and lower computing efficiency when segmenting clothing images.

[0003] In video multi-frame scenarios such as image segmentation of washing machine clothing or synchronous segmentation scenarios of multiple clothing images, since the clothing in the washing machine is entangled and covered with each other, it is impossible to effectively observe the main features of each piece of clothing from a single clothing image. It is necessary to record a clothing video during the rolling process of the inner barrel of the washing machine for video detection. However, existing clothing segmentation methods usually segment a single clothing image and cannot effectively handle video multi-frame scenarios and synchronous segmentation scenarios of multiple clothing images, and it is difficult to respond to various clothing image segmentation scenarios in real time, resulting in lower accuracy and poorer clothing segmentation effect. Summary of the Invention

[0004] The main purpose of the embodiments of the present application is to propose a clothing segmentation method, device, electronic device, and storage medium, which can realize automatic segmentation of clothing images and improve the efficiency and accuracy of clothing segmentation.

[0005] On the one hand, the embodiments of the present application propose a clothing segmentation method, and the method includes the following steps:

[0006] Obtain a clothing video to be detected;

[0007] Perform image screening on the clothing video to be detected to determine multiple clothing images to be segmented;

[0008] Synchronously input the multiple clothing images to be segmented into a clothing segmentation model, and use the clothing segmentation model to output clothing segmentation results corresponding to each of the clothing images to be segmented.

[0009] In some embodiments, the performing image screening on the clothing video to be detected to determine multiple clothing images to be segmented specifically includes:

[0010] Perform frame image division on the clothing video to be detected to obtain multiple video frame clothing images;

[0011] Perform image quality screening on each of the video frame clothing images to determine corresponding multiple first screened images;

[0012] Perform image similarity screening on multiple of the first screened images to determine corresponding multiple second screened images, and determine each of the second screened images as the clothing image to be segmented.

[0013] In some embodiments, the method further includes:

[0014] Construct the clothing segmentation model;

[0015] Obtain a sample data set; the sample data set includes multiple clothing sample image sequences with clothing segmentation labels;

[0016] Use the sample data set to train and optimize the clothing segmentation model.

[0017] In some embodiments, the clothing segmentation model includes a feature extraction module, a feature fusion module, a feature separation module, and a clothing segmentation module. Synchronously inputting multiple of the clothing images to be segmented into the clothing segmentation model and using the clothing segmentation model to output the clothing segmentation results corresponding to each of the clothing images to be segmented specifically includes:

[0018] Input multiple of the clothing images to be segmented into the feature extraction module, and use the feature extraction module to extract the image features corresponding to each of the clothing images to be segmented;

[0019] Synchronously superimpose and input the image features corresponding to each of the clothing images to be segmented into the feature fusion module, and use the feature fusion module to perform feature fusion and output image fusion features;

[0020] Input the image fusion features into the feature separation module, and use the feature separation module to output corresponding multiple image separation features;

[0021] Synchronously input multiple of the image separation features and each of the clothing images to be segmented into the clothing segmentation module, and use the clothing segmentation module to perform clothing segmentation and synchronously output the clothing segmentation results corresponding to each of the clothing images to be segmented.

[0022] In some embodiments, the performing image quality screening on each of the video frame clothing images to determine corresponding multiple first screened images specifically includes:

[0023] Obtain an image quality evaluation model;

[0024] Use the image quality evaluation model to perform image quality scoring on each of the video frame clothing images to determine the image quality scores corresponding to each of the video frame clothing images;

[0025] Rank the scores from high to low according to the image quality scores corresponding to each of the video frame clothing images, and screen out a plurality of the first screened images corresponding to a preset ranking threshold, where the first screened image is the video frame clothing image whose current ranking is higher than the preset ranking threshold.

[0026] In some embodiments, the method of performing image similarity screening on the plurality of the first screened images to determine corresponding second screened images and determining each of the second screened images as the clothing image to be segmented specifically includes:

[0027] Obtain any one of the first screened images as the current comparison image;

[0028] Perform image similarity comparison on the current comparison image and each of the remaining first screened images in sequence to determine the image similarity corresponding to the current comparison image and each of the remaining second screened images;

[0029] According to each of the image similarities, determine whether there are several target similar images among the remaining plurality of the first screened images, where the image similarity between the target similar image and the current comparison image is greater than a preset similarity threshold;

[0030] When it is determined that there are no several target similar images, determine the current comparison image as the second screened image;

[0031] When it is determined that there are several target similar images, obtain each first image quality score, and according to each of the first image quality scores, determine a target screened image from the several target similar images, and add all the remaining target similar images except the target screened image to a pre-constructed image filtering data set, where the first image quality score is the image quality score corresponding to the target similar image, and the target screened image is the target similar image with the highest image quality score;

[0032] Obtain a second image quality score, and compare the second image quality score with the image quality score corresponding to the target screened image, where the second image quality score is the image quality score corresponding to the current comparison image;

[0033] When the second image quality score is greater than the image quality score corresponding to the target screened image, determine the current comparison image as the second screened image, and add the target screened image to the image filtering data set, otherwise, determine the target screened image as the second screened image, and add the current comparison image to the image filtering data set;

[0034] Determine whether there is the first screened image that has not been added to the image filtering dataset. When it is determined that there is the first screened image that has not been added to the image filtering dataset, obtain any first screened image that has not been added to the image filtering dataset as the current comparison image, and then return to the step of sequentially performing image similarity comparison on the current comparison image and each of the remaining first screened images to determine the image similarity corresponding to the current comparison image and each of the remaining second screened images, until it is determined that there is no first screened image that has not been added to the image filtering dataset.

[0035] In some embodiments, the clothing segmentation module includes a plurality of clothing segmentation units. Synchronously inputting the multiple image separation features and each to-be-segmented clothing image into the clothing segmentation module, and using the clothing segmentation module to perform clothing segmentation and synchronously output the clothing segmentation results corresponding to each to-be-segmented clothing image specifically includes:

[0036] Determine the target image segmentation unit corresponding to each to-be-segmented clothing image from the multiple clothing segmentation units;

[0037] For each to-be-segmented clothing image, input the multiple image separation features and the to-be-segmented clothing image into the target image segmentation unit, and use the target image segmentation unit to output the clothing segmentation result corresponding to the to-be-segmented clothing image.

[0038] On the other hand, an embodiment of the present application proposes a clothing segmentation device, and the device includes:

[0039] A first module, configured to obtain a to-be-detected clothing video;

[0040] A second module, configured to perform image screening on the to-be-detected clothing video to determine a plurality of to-be-segmented clothing images;

[0041] A third module, configured to synchronously input the plurality of to-be-segmented clothing images into a clothing segmentation model, and use the clothing segmentation model to output the clothing segmentation results corresponding to each to-be-segmented clothing image.

[0042] On the other hand, an embodiment of the present application proposes an electronic device, and the electronic device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the foregoing clothing segmentation method is implemented.

[0043] On the other hand, an embodiment of the present application proposes a computer-readable storage medium, and the computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the foregoing clothing segmentation method is implemented.

[0044] The embodiments of the present application at least include the following beneficial effects: A clothing segmentation method, device, electronic device, and storage medium provided by the present application obtain a video of clothing to be detected, perform image screening on the video of clothing to be detected, determine multiple clothing images to be segmented, synchronously input the multiple clothing images to be segmented into a clothing segmentation model, and use the clothing segmentation model to output clothing segmentation results corresponding to each clothing image to be segmented. The present application can effectively handle video multi-frame scenarios and synchronous segmentation scenarios of multiple clothing images, realize automatic segmentation of clothing images, be applicable to various clothing image segmentation scenarios, improve the efficiency and accuracy of clothing segmentation, and enhance the clothing segmentation effect. Brief Description of the Drawings

[0045] Figure 1 is a flowchart of a clothing segmentation method provided by an embodiment of the present application;

[0046] Figure 2 is a schematic structural diagram of a clothing segmentation model in an embodiment of the present application;

[0047] Figure 3 is a schematic structural diagram of a clothing segmentation device provided by an embodiment of the present application;

[0048] Figure 4 is a schematic hardware structure diagram of an electronic device provided by an embodiment of the present application. Detailed Description of the Embodiments

[0049] In order to make the objectives, technical solutions, and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application. They are only examples of devices and methods consistent with some aspects of the embodiments of the present application detailed in the appended claims.

[0050] It can be understood that the terms "first", "second", etc. used in the present application may be used herein to describe various concepts, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if", "when" as used herein may be interpreted as "when...", "while...", or "in response to determining".

[0051] The terms "at least one", "a plurality", "each", "any one", etc. used in this application, at least one includes one, two or more than two, a plurality includes two or more than two, each refers to each of the corresponding plurality, and any one refers to any one of the plurality.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0053] Refer to Figure 1 , Figure 1 is an optional flowchart of a clothing segmentation method provided by an embodiment of this application. The method may include but is not limited to steps S101 to S103:

[0054] Step S101, obtain the video of the clothing to be detected;

[0055] Step S102, perform image screening on the video of the clothing to be detected to determine a plurality of clothing images to be segmented;

[0056] Step S103, synchronously input the plurality of clothing images to be segmented into the clothing segmentation model, and use the clothing segmentation model to output the clothing segmentation results corresponding to each clothing image to be segmented.

[0057] In some embodiments, dynamically collect the video of the clothing to be detected, perform video frame division and image screening on the video of the clothing to be detected to obtain the corresponding plurality of clothing images to be segmented. The image screening may include but is not limited to image quality evaluation and image similarity screening.

[0058] In some embodiments, step S102 may include but is not limited to steps S201 to S203:

[0059] Step S201, perform frame image division on the video of the clothing to be detected to obtain a plurality of video frame clothing images;

[0060] Step S202, perform image quality screening on each video frame clothing image to determine the corresponding plurality of first screened images;

[0061] Step S203, perform image similarity screening on the plurality of first screened images to determine the corresponding plurality of second screened images, and determine each second screened image as the clothing image to be segmented.

[0062] In some embodiments, an image quality evaluation model is constructed, and the image quality of each video frame clothing image is screened using the image quality evaluation model. Optionally, the image quality evaluation model is a deep learning model. An image quality evaluation sample set is obtained, and the image quality evaluation model is trained using the image quality evaluation sample set. The image quality evaluation sample set includes multiple clothing images with image quality score labels, and the image quality score labels may include, but are not limited to, image quality scores, image sharpness, and image sharpening degree, etc.

[0063] In some embodiments, step S103 may include, but is not limited to, steps S301 to S304:

[0064] Step S301: Input multiple clothing images to be segmented into the feature extraction module, and use the feature extraction module to extract the image features corresponding to each clothing image to be segmented.

[0065] Step S302: Synchronously stack and input the image features corresponding to each clothing image to be segmented into the feature fusion module, and use the feature fusion module to perform feature fusion and output image fusion features.

[0066] Step S303: Input the image fusion features into the feature separation module, and use the feature separation module to output corresponding multiple image separation features.

[0067] Step S304: Synchronously input the multiple image separation features and each clothing image to be segmented into the clothing segmentation module, and use the clothing segmentation module to perform clothing segmentation and synchronously output the clothing segmentation results corresponding to each clothing image to be segmented.

[0068] In some embodiments, referring to Figure 2 , Figure 2It is an optional structural schematic diagram of the clothing segmentation model in the embodiments of the present application. The clothing images to be segmented from 1 to N are synchronously input into the clothing segmentation model. The clothing segmentation model includes a feature extraction module, a feature fusion module, a feature separation module, and a clothing segmentation module. The feature extraction module is constructed based on CNN (Convolutional Neural Networks), and is used to extract image features. The feature extraction module includes feature extraction units from 1 to N. Each feature extraction unit is used to extract features from the corresponding clothing images to be segmented, and determine the image features of each clothing image to be segmented. Then, the image features of each clothing image to be segmented are treated equally and synchronously input into the feature fusion module to complete feature superposition and fusion. The feature fusion module is a high-level semantic feature fusion module constructed based on CNN, and is used to further fuse the image features of each clothing image to be segmented output by the feature extraction module to obtain image fusion features. The feature separation module is used to perform feature separation so that the subsequent clothing segmentation module can more effectively learn the features of a single image. The clothing segmentation module is used to combine multiple image separation features and synchronously perform image segmentation on each clothing image to be segmented. The clothing segmentation module includes image segmentation units from 1 to N. Each image segmentation unit is used to perform clothing segmentation on the corresponding clothing image to be segmented. The structure of the image segmentation unit is the instance segmentation Head structure of YOLOv8.

[0069] In some embodiments, optionally, a clothing segmentation model is constructed, and a sample data set is obtained. The sample data set includes multiple clothing sample image sequences with clothing segmentation labels, and the clothing segmentation model is trained and optimized using the sample data set.

[0070] In step S304 of some embodiments, the target image segmentation unit corresponding to each clothing image to be segmented is determined from multiple clothing segmentation units. For each clothing image to be segmented, multiple image separation features and the clothing image to be segmented are input into the target image segmentation unit, and the clothing segmentation result corresponding to the clothing image to be segmented is output using the target image segmentation unit.

[0071] Optionally, as Figure 2 shown, assuming there are clothing images to be segmented from 1 to N, the clothing segmentation module includes clothing segmentation units from 1 to N. The clothing segmentation unit 1 is used to perform image segmentation on the clothing image to be segmented 1 according to multiple picture separation features, and output the clothing segmentation result corresponding to the clothing image to be segmented 1. The clothing segmentation unit 2 is used to perform image segmentation on the clothing image to be segmented 2 according to multiple picture separation features, and output the clothing segmentation result corresponding to the clothing image to be segmented 2, and so on.

[0072] In some embodiments, step S202 may include but is not limited to steps S401 to S403:

[0073] Step S401, obtain an image quality evaluation model;

[0074] Step S402, use the image quality evaluation model to perform image quality scoring on each video frame clothing image, and determine the image quality score corresponding to each video frame clothing image;

[0075] Step S403, according to the image quality scores corresponding to each video frame clothing image, sort the scores from high to low, and screen out multiple first screened images corresponding to a preset sorting threshold, where the first screened image is a video frame clothing image whose current sorting is higher than the preset sorting threshold.

[0076] In some embodiments, according to the image quality scores corresponding to each video frame clothing image, the top M images with high image quality scores are screened out from multiple video frame clothing images, M is the preset sorting threshold. Exemplarily, assume there are video frame clothing images 1 to N, and M is 15. Use the image quality evaluation model to output the image quality scores corresponding to each video frame clothing image, sort the scores from high to low, and screen out the top 15 images with high image quality scores from N video frame clothing images as the first screened images.

[0077] In some embodiments, step S203 may include but is not limited to steps S501 to S508:

[0078] Step S501, obtain any one of the first screened images as the current comparison image;

[0079] Step S502, sequentially perform image similarity comparison between the current comparison image and the remaining first screened images, and determine the image similarity corresponding to the current comparison image and the remaining second screened images;

[0080] Step S503, according to each image similarity, determine whether there are several target similar images among the remaining multiple first screened images, and the image similarity between the target similar image and the current comparison image is greater than the preset similarity threshold;

[0081] Step S504, when it is determined that there are no several target similar images, determine the current comparison image as the second screened image;

[0082] Step S505, when it is determined that there are several target similar images, obtain each first image quality score, and according to each first image quality score, determine a target screened image from the several target similar images, and add all the remaining target similar images except the target screened image to the pre-constructed image filtering dataset, where the first image quality score is the image quality score corresponding to the target similar image, and the target screened image is the target similar image with the highest image quality score;

[0083] Step S506: Obtain the second image quality score, and compare the second image quality score with the image quality score corresponding to the target screening image, where the second image quality score is the image quality score corresponding to the current comparison image.

[0084] Step S507: When the second image quality score is greater than the image quality score corresponding to the target screening image, determine the current comparison image as the second screening image, and add the target screening image to the image filtering dataset; otherwise, determine the target screening image as the second screening image, and add the current comparison image to the image filtering dataset.

[0085] Step S508: Determine whether there is a first screening image that has not been added to the image filtering dataset. When it is determined that there is a first screening image that has not been added to the image filtering dataset, obtain any first screening image that has not been added to the image filtering dataset as the current comparison image, and then return to the step of sequentially performing image similarity comparison between the current comparison image and each of the remaining first screening images to determine the image similarity between the current comparison image and each of the corresponding second screening images until it is determined that there is no first screening image that has not been added to the image filtering dataset.

[0086] In some embodiments, perform image similarity screening on multiple first screening images, and screen out multiple first screening images with large image differences as the second screening images. Optionally, assume there are first screening images 1 to M. First, select the first screening image 1 as the current comparison image, and compare the first screening image 1 with the first screening images 2 to M in sequence to determine the image similarity between the first screening image 1 and each of the remaining first screening images, and obtain the preset image similarity threshold as Q s , after image similarity judgment, the image similarity between the first screening image 1 and the first screening image 2 is Q 1 >Q s , the image similarity between the first screening image 1 and the first screening image 3 is Q 2 >Q s , therefore, determine both the first screening image 2 and the first screening image 3 as the target similar images of the current comparison image, and obtain the image quality score S of the current comparison image 1 , the image quality score S of the first screening image 2 2 and the image quality score S of the first screening image 3 3 , where, assume S 3 >S 2 , then add the first screening image 2 to the image filtering dataset, and determine the first screening image 3 as the target screening image. Assume S 3 >S 1, that is, the image quality score of the target screening image is greater than that of the current comparison image. Add the current comparison image to the image filtering dataset, and determine the first screening image 3 as the second screening image. At this time, the first screening images not added to the image filtering dataset include the first screening image 4 to the first screening image M. Then, select any first screening image from the remaining first screening images 4 to the first screening image M as the current comparison image, and perform the next round of image similarity screening until all the first screening images are added to the image filtering dataset or determined as the second screening image, and then exit the image similarity screening.

[0087] Refer to Figure 3 , Figure 3 FIG. is an optional structural schematic diagram of a clothing segmentation device provided by an embodiment of the present application. The device is used to implement the above clothing segmentation method, and the device may include:

[0088] The first module is used to obtain the video of the clothing to be detected;

[0089] The second module is used to perform image screening on the video of the clothing to be detected and determine multiple clothing images to be segmented;

[0090] The third module is used to synchronously input multiple clothing images to be segmented into the clothing segmentation model, and use the clothing segmentation model to output the clothing segmentation results corresponding to each clothing image to be segmented.

[0091] It can be understood that the content in the above method embodiments is applicable to the device embodiments of the present application. The functions specifically implemented by the device embodiments of the present application are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those of the above method embodiments.

[0092] An embodiment of the present application further provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above clothing segmentation method is implemented. The electronic device can be any intelligent terminal including a tablet computer, etc.

[0093] It can be understood that the content in the above method embodiments is applicable to the device embodiments of the present application. The functions specifically implemented by the device embodiments of the present application are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those of the above method embodiments.

[0094] Please refer to Figure 4 , Figure 4 FIG. schematically shows the hardware structure of an electronic device according to another embodiment. The electronic device includes:

[0095] The processor 901 can be implemented in the form of a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0096] The memory 902 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 902 can store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 902 and are called by the processor 901 to execute the clothing segmentation method in the embodiments of the present application;

[0097] The input / output interface 903 is used to implement information input and output;

[0098] The communication interface 904 is used to implement communication interaction between this device and other devices, and can implement communication through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.);

[0099] The bus 905 transmits information between various components of the device (such as the processor 901, the memory 902, the input / output interface 903, and the communication interface 904);

[0100] Among them, the processor 901, the memory 902, the input / output interface 903, and the communication interface 904 achieve communication connections with each other inside the device through the bus 905.

[0101] The embodiments of the present application also provide a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned clothing segmentation method is implemented.

[0102] It can be understood that the content in the above method embodiments is applicable to the embodiments of this storage medium. The functions specifically implemented by the embodiments of this storage medium are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those in the above method embodiments.

[0103] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0104] A clothing segmentation method, apparatus, electronic device, and storage medium provided by an embodiment of the present application can effectively handle the synchronous segmentation of video multi-frame scenarios and multiple clothing images, realize automatic segmentation of clothing images, improve the efficiency and accuracy of clothing segmentation, and enhance the clothing segmentation effect.

[0105] The embodiments described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art will know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present application are equally applicable to similar technical problems.

[0106] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than those shown in the figures, or combine certain steps, or different steps.

[0107] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0108] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and appropriate combinations thereof.

[0109] In the description of the present application and the above-mentioned drawings, terms such as "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that comprises a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0110] It should be understood that in the present application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects and indicates that three relationships may exist. For example, "A and / or B" may mean: only A exists, only B exists, and both A and B exist simultaneously. Among them, A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one (one) of the following" or its similar expression means any combination of these items, including any combination of single item (one) or plural items (ones). For example, at least one (one) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0111] In several embodiments provided by the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the above-mentioned division of units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces, and the indirect coupling or communication connection of devices or units can be in electrical, mechanical, or other forms.

[0112] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0113] In addition, in each embodiment of the present application, each functional unit can be integrated into one processing unit, can exist physically alone for each unit, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0114] It should be recognized that the embodiments of the present invention can be implemented or carried out by computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer-readable memory. The method can be implemented in a computer program using standard programming techniques including a non-transitory computer-readable storage medium configured with the computer program, wherein the storage medium so configured causes the computer to operate in a specific and predefined manner - according to the methods and drawings described in the specific embodiments. Each program can be implemented in a high-level procedural or object-oriented programming language to communicate with the computer system. However, if desired, the program can be implemented in assembly or machine language. In any case, the language can be a compiled or interpreted language. In addition, for this purpose the program can run on a dedicated integrated circuit programmed for this purpose.

[0115] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of the present application. The foregoing storage medium includes: various media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc that can store programs.

[0116] The preferred embodiments of the embodiments of the present application have been described above with reference to the accompanying drawings. This does not limit the scope of the rights of the embodiments of the present application. Any modification, equivalent replacement, and improvement made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the rights of the embodiments of the present application.

Claims

1. A clothing segmentation method, characterized in that: The method comprises the following steps: Get the video of the clothing to be tested; Performing image screening on the clothing video to be detected to determine a plurality of clothing images to be segmented; The plurality of clothing images to be segmented are synchronously input into a clothing segmentation model, and the clothing segmentation model is used to output clothing segmentation results corresponding to each of the clothing images to be segmented.

2. The clothing segmentation method according to claim 1, characterized in that: The performing image screening on the clothing video to be detected to determine a plurality of clothing images to be segmented specifically includes: Dividing the to-be-detected clothing video into frame images to obtain a plurality of video frame clothing images; Performing image quality screening on the clothing images of each of the video frames to determine a corresponding plurality of first screening images; Perform image similarity screening on the plurality of first screening images, determine a plurality of corresponding second screening images, and determine each of the second screening images as the clothing image to be segmented.

3. The clothing segmentation method according to claim 1, characterized in that: The method further comprises: Constructing the clothing segmentation model; Acquire a sample data set; the sample data set includes a plurality of clothing sample image sequences with clothing segmentation labels; The clothing segmentation model is trained and optimized using the sample data set.

4. The clothing segmentation method according to claim 1, characterized in that: The clothing segmentation model comprises a feature extraction module, a feature fusion module, a feature separation module and a clothing segmentation module. The multiple clothing images to be segmented are synchronously input into the clothing segmentation model, and the clothing segmentation model is used to output clothing segmentation results corresponding to each of the clothing images to be segmented, specifically comprising: Inputting a plurality of clothing images to be segmented into the feature extraction module, and using the feature extraction module to extract image features corresponding to each of the clothing images to be segmented; The image features corresponding to each of the clothing images to be segmented are synchronously superimposed and input into the feature fusion module, and the feature fusion module is used to perform feature fusion and output image fusion features; Inputting the image fusion features into the feature separation module, and using the feature separation module to output corresponding multiple image separation features; The plurality of image separation features and each of the clothing images to be segmented are synchronously input into the clothing segmentation module, clothing segmentation is performed using the clothing segmentation module, and the clothing segmentation results corresponding to each of the clothing images to be segmented are synchronously output.

5. The clothing segmentation method according to claim 2, characterized in that: The performing image quality screening on the clothing images of each of the video frames to determine a plurality of corresponding first screening images specifically includes: Obtain an image quality evaluation model; Using the image quality evaluation model to score the image quality of the clothing image in each video frame, and determine the image quality score corresponding to the clothing image in each video frame; According to the image quality scores corresponding to the clothing images of each video frame, the scores are sorted from high to low, and a plurality of the first filtered images corresponding to a preset sorting threshold are screened out, wherein the first filtered images are the clothing images of the video frames whose current sorting is higher than the preset sorting threshold.

6. The clothing segmentation method according to claim 5, characterized in that: The performing image similarity screening on the plurality of first screening images to determine a plurality of corresponding second screening images, and determining each of the second screening images as the clothing image to be segmented specifically includes: Acquire any one of the first screening images as a current comparison image; performing image similarity comparison on the current comparison image and the remaining first screening images in sequence, and determining image similarities between the current comparison image and the remaining second screening images; According to the image similarities, determining whether there are a plurality of target similar images from the remaining plurality of the first screening images, wherein the image similarity between the target similar image and the image corresponding to the current comparison image is greater than a preset similarity threshold; When it is determined that there are no multiple target similar images, determining the current comparison image as the second screening image; When it is determined that there are several target similar images, obtaining each first image quality score, determining a target screening image from the several target similar images according to each first image quality score, and adding all the target similar images except the target screening image to the pre-constructed image filtering data set, wherein the first image quality score is the image quality score corresponding to the target similar image, and the target screening image is the target similar image with the highest image quality score; Acquire a second image quality score, and compare the second image quality score with the image quality score corresponding to the target screening image, wherein the second image quality score is the image quality score corresponding to the current comparison image; When the second image quality score is greater than the image quality score corresponding to the target screening image, the current comparison image is determined as the second screening image, and the target screening image is added to the image filtering data set; otherwise, the target screening image is determined as the second screening image, and the current comparison image is added to the image filtering data set; Determine whether there is the first screening image that has not been added to the image filtering data set. When it is determined that there is the first screening image that has not been added to the image filtering data set, obtain any one of the first screening images that has not been added to the image filtering data set as the current comparison image, and then return to the step of performing image similarity comparison on the current comparison image and the remaining first screening images in sequence to determine the image similarity between the current comparison image and the remaining second screening images, until it is determined that there is no first screening image that has not been added to the image filtering data set.

7. The clothing segmentation method according to claim 4, characterized in that: The clothing segmentation module comprises a plurality of clothing segmentation units, wherein the plurality of image separation features and each of the clothing images to be segmented are synchronously input into the clothing segmentation module, clothing segmentation is performed using the clothing segmentation module, and the clothing segmentation results corresponding to each of the clothing images to be segmented are synchronously output, specifically comprising: Determining a target image segmentation unit corresponding to each of the to-be-segmented clothing images from the plurality of clothing segmentation units; For each of the clothing images to be segmented, the plurality of image separation features and the clothing image to be segmented are input into the target image segmentation unit, and the target image segmentation unit is used to output the clothing segmentation result corresponding to the clothing image to be segmented.

8. A clothing dividing device, characterized in that: The device comprises: The first module is used to obtain the video of the clothing to be detected; The second module is used to screen the image of the clothing video to be detected and determine a plurality of clothing images to be segmented; The third module is used to synchronously input the multiple clothing images to be segmented into the clothing segmentation model, and use the clothing segmentation model to output clothing segmentation results corresponding to each of the clothing images to be segmented.

9. An electronic device, characterized in that: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the clothing segmentation method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the clothing segmentation method according to any one of claims 1 to 7 is implemented.

Images