Model generation method and device, image processing method and device, medium and electronic equipment

Abstract

The embodiment of the invention relates to a model generation method and device, an image processing method and device, a medium and electronic equipment. The method comprises the steps of obtaining first data, generating a first sample set according to the first data, and training a preset model according to the first sample set to obtain a target model. Wherein the first data comprises first text data, the first sample set comprises a first sample image and a first sample label corresponding to the first sample image, the first sample label is used for labeling sample object information corresponding to the first sample image, and the target model can be used for carrying out image segmentation on a first target image; and obtaining target object information corresponding to the first target image, wherein the target object information comprises an object type of the target object and / or position information of the target object in the first target image.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular, to a model generation method, an image processing method, a device, a medium and an electronic device. Background Art

[0002] With the advancement of computer technology, image segmentation has become an important research direction in the field of computer vision. Building a target model based on machine learning and performing image segmentation through the target model is a widely used image segmentation method.

[0003] In order to train the target model, sample images need to be collected manually and annotated manually, which is inefficient. Summary of the invention

[0004] This summary is provided to introduce concepts in a brief form that will be described in detail in the detailed description below. This summary is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0005] According to a first aspect of an embodiment of the present disclosure, a model generation method is provided, the method comprising:

[0006] Acquire first data; wherein the first data includes first text data;

[0007] Generate a first sample set according to the first data; wherein the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to mark sample object information corresponding to the first sample image;

[0008] The preset model is trained according to the first sample set to obtain a target model; wherein the target model is used to perform image segmentation on the first target image to obtain target object information corresponding to the first target image, and the target object information includes the object type of the target object and / or the location information of the target object in the first target image.

[0009] In some embodiments, generating a first sample set according to the first data includes:

[0010] Acquire second text data according to the first text data;

[0011] The first sample set is generated according to the second text data.

[0012] In some embodiments, acquiring the second text data according to the first text data includes:

[0013] Performing prompt text expansion on the first text data to obtain second text data; or,

[0014] The first text data is used as the second text data.

[0015] In some embodiments, generating the first sample set according to the second text data includes:

[0016] generating a first sample image according to the second text data;

[0017] The first sample image is labeled by using a pre-generated first model to obtain a first sample label corresponding to the first sample image; wherein the first model is a pre-trained vertical classification model, or the first model is a model obtained based on small sample segmentation training.

[0018] In some embodiments, generating a first sample image according to the second text data includes:

[0019] According to the second text data, the first sample image that meets the first search condition is searched in the pre-generated first data set; wherein the first data set includes a plurality of sample images, and the first search condition is used to characterize that the similarity between the first image feature of the first sample image and the text feature of the second text data is greater than or equal to a first preset similarity threshold; or

[0020] The second text data is input into a pre-generated first text image generation model to obtain a first sample image output by the first text image generation model.

[0021] In some embodiments, the first data further includes first image data; and generating a first sample image according to the second text data includes:

[0022] According to the second text data and / or the first image data, a search is performed in a pre-generated first data set to obtain the first sample image that meets a second search condition; wherein the first data set includes a plurality of sample images, and the second search condition is used to characterize that the similarity between the first image feature of the first sample image and the text feature of the second text data is greater than or equal to a second preset similarity threshold, or that the similarity between the first image feature of the first sample image and the second image feature of the first image data is greater than or equal to a third preset similarity threshold; or,

[0023] The second text data and the first image data are input into a pre-generated diffusion model to obtain a first sample image output by the diffusion model.

[0024] In some embodiments, generating the first sample set according to the second text data includes:

[0025] The second text data is input into a pre-generated second text image generation model to obtain a first sample image output by the second text image generation model and a first sample label corresponding to the first sample image.

[0026] In some embodiments, the first data further includes unlabeled first image data; and generating the first sample set according to the second text data includes:

[0027] According to the second text data and / or the first image data, a search is performed in a pre-generated second data set to obtain the first sample image that meets a third search condition and a first sample label corresponding to the first sample image; wherein the second data set includes a plurality of labeled sample images, and the third search condition is used to characterize that the similarity between the first image feature of the first sample image and the image feature of the retrieved sample image is greater than or equal to a fourth preset similarity threshold.

[0028] In some embodiments, the first data further includes second image data with a label; and generating the first sample set according to the second text data includes:

[0029] According to the second image data, the second text data and the preset sample label, a plurality of first sample images are generated based on a control network; wherein the plurality of first sample images correspond to the same preset sample label.

[0030] In some embodiments, the method further comprises:

[0031] The trained target model is tested to obtain a model test index of the target model.

[0032] According to a second aspect of an embodiment of the present disclosure, there is provided an image processing method, the method comprising:

[0033] Acquire a first target image to be processed;

[0034] Inputting the first target image into a pre-generated target model, performing image segmentation on the first target image, and obtaining target object information corresponding to the first target image, wherein the target object information includes an object type of the target object and / or position information of the target object in the first target image;

[0035] Among them, the target model is a model obtained by training a preset model according to a first sample set, the first sample set is a sample set generated according to first data, the first data includes first text data, the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to mark the sample object information corresponding to the first sample image.

[0036] In some embodiments, the target model is a model generated according to the method described in the first aspect of the present disclosure.

[0037] According to a third aspect of an embodiment of the present disclosure, a model generation device is provided, the device comprising:

[0038] A first acquisition module, used to acquire first data; wherein the first data includes first text data;

[0039] A sample generation module, configured to generate a first sample set according to the first data; wherein the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to mark sample object information corresponding to the first sample image;

[0040] A model generation module is used to train a preset model according to the first sample set to obtain a target model; wherein the target model is used to perform image segmentation on a first target image to obtain target object information corresponding to the first target image, and the target object information includes the object type of the target object and / or the location information of the target object in the first target image.

[0041] According to a fourth aspect of an embodiment of the present disclosure, there is provided an image processing apparatus, the apparatus comprising:

[0042] A second acquisition module, used for acquiring a first target image to be processed;

[0043] an image processing module, configured to input the first target image into a pre-generated target model, perform image segmentation on the first target image, and obtain target object information corresponding to the first target image, wherein the target object information includes an object type of the target object and / or position information of the target object in the first target image;

[0044] Among them, the target model is a model obtained by training a preset model according to a first sample set, the first sample set is a sample set generated according to first data, the first data includes first text data, the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to mark the sample object information corresponding to the first sample image.

[0045] According to a fifth aspect of an embodiment of the present disclosure, a computer-readable medium is provided, on which a computer program is stored. When the computer program is executed by a processing device, the steps of the method described in the first aspect or the second aspect of the present disclosure are implemented.

[0046] According to a sixth aspect of an embodiment of the present disclosure, there is provided an electronic device, including:

[0047] a storage device having a computer program stored thereon;

[0048] A processing device is used to execute the computer program in the storage device to implement the steps of the method described in the first aspect or the second aspect of the present disclosure.

[0049] By adopting the above technical solution, the first data is obtained, a first sample set is generated according to the first data, and a preset model is trained according to the first sample set to obtain a target model. The first data includes first text data, the first sample set includes a first sample image and a first sample label corresponding to the first sample image, the first sample label is used to mark the sample object information corresponding to the first sample image, and the target model can be used to perform image segmentation on the first target image to obtain the target object information corresponding to the first target image, and the target object information includes the object type of the target object and / or the location information of the target object in the first target image. In this way, a sample set for model training can be efficiently obtained, which improves the efficiency of model generation.

[0050] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale. In the drawings:

[0052] Figure 1 The figure is a flow chart of a model generation method according to an embodiment of the present disclosure.

[0053] Figure 2 is based on Figure 1 The illustrated embodiment shows a flow chart of step S102.

[0054] Figure 3 is based on Figure 2 The illustrated embodiment shows a flow chart of step S1022.

[0055] Figure 4 is based on Figure 2 The illustrated embodiment shows another flow chart of step S1022.

[0056] Figure 5 is based on Figure 2 The illustrated embodiment shows another flow chart of step S1022.

[0057] Figure 6 is based on Figure 1 The illustrated embodiment shows a flowchart of another step S1022.

[0058] Figure 7 The figure is a flowchart of a model updating method according to an embodiment of the present disclosure.

[0059] Figure 8 It is a flowchart of an image processing method according to an embodiment of the present disclosure.

[0060] Fig. 9 It is a block diagram of a model generation device according to an embodiment of the present disclosure.

[0061] Fig.10 It is a block diagram of an image processing device according to an embodiment of the present disclosure.

[0062] Fig.11 It is a block diagram of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0063] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0064] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0065] The term "including" and its variations used in the present disclosure are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0066] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0067] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise clearly indicated in the context, they should be understood as "one or more". In the description of the present disclosure, unless otherwise specified, "multiple" refers to two or more than two, and other quantifiers are similar; "at least one item", "one or more items" or similar expressions refer to any combination of these items, including any combination of single items or plural items. For example, at least one item a can represent any number of a; for another example, one or more items of a, b and c can represent: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or multiple; "and / or" is a kind of association relationship that describes the associated objects, indicating that there can be three relationships, for example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The character “ / ” indicates that the preceding and following objects are in an “or” relationship. The singular forms “a”, “an”, “an”, “said” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise.

[0068] Although operations or steps are described in a specific order in the drawings in the embodiments of the present disclosure, it should not be understood that it is required to perform these operations or steps in the specific order shown or in a serial order, or to perform all the operations or steps shown to obtain the desired results. In the embodiments of the present disclosure, these operations or steps can be performed in series; these operations or steps can also be performed in parallel; or some of these operations or steps can be performed.

[0069] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0070] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0071] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0072] As an optional but non-limiting implementation, in response to receiving an active request from the user, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0073] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet the relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0074] At the same time, it is understandable that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and relevant provisions.

[0075] The present disclosure is described below in conjunction with specific embodiments.

[0076] Figure 1 is a flow chart of a model generation method according to an embodiment of the present disclosure. The method can be applied to electronic devices, which may include terminal devices, such as smart phones, smart wearable devices, smart speakers, smart tablets, PDAs (Personal Digital Assistants), CPEs (Customer Premise Equipment), personal computers, vehicle-mounted terminals, etc.; the electronic devices may also include servers, such as local servers or cloud servers. Figure 1 As shown, the method may include:

[0077] S101: Obtain first data.

[0078] In some embodiments, the first data includes first text data.

[0079] In some embodiments, the first data may be multimodal data. For example, the first data may include at least one of first text data, unlabeled first image data, and labeled second image data.

[0080] For example, the first data may include first text data and first image data without a label.

[0081] For another example, the first data may include first text data and second image data with a label.

[0082] For another example, the first data may include first text data, unlabeled first image data, and labeled second image data.

[0083] The first text data may be text information describing the sample object. For example, the first text data may describe the object type of the sample object (e.g., portrait, cat, dog, etc.), or may describe the expected position area of ​​the sample object (e.g., lower left corner, lower right corner, middle 1 / 2 area, etc.). The first text data may be used to generate a first sample image.

[0084] The first image data may include a picture or a video, for example, a picture or a video providing a business scenario.

[0085] In one implementation, the first image data may be used to generate a first sample image.

[0086] In another implementation, the first image data may be intelligently annotated based on an existing platform or model to obtain labeled first image data.

[0087] The second image data may be labeled image data, for example, may be pair data, which may include an original image and a label, wherein the label may be correct annotation information GT, and the label may include relevant information of a sample object in the second image data, wherein the relevant information may include object type and location information, and the location information may be represented in the form of a mask MASK.

[0088] It should be noted that the first data may be data input by a user. For example, the user's input may be received and the first data may be obtained. The first data may also be data obtained from a database. For example, the first data may be downloaded from the database. The first data may also be data determined from a database based on a user's selection. For example, a database may provide a user with multiple data so that the user may select and determine the first data from the multiple data.

[0089] S102: Generate a first sample set according to first data.

[0090] The first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to mark sample object information corresponding to the first sample image.

[0091] Optionally, the sample object information may include an object type of the sample object and / or position information of the sample object in the first sample image.

[0092] It should be noted that the sample object can be any object, animal, plant, portrait, human body part, animal or plant part, animal or plant variety, etc., and the embodiments of the present disclosure are not limited to this. The object type can be the type of the sample object. For example, the object type can include animals, plants, objects or people, etc. For another example, the object type can include subdivided types under animals (such as cats, dogs, sheep, etc.). For another example, the object type can also include more detailed types (such as white cats, black cats, yellow cats, etc.), and the embodiments of the present disclosure are not limited to this. The position information can be used to indicate the area of ​​the sample object in the first sample image. For example, the position information can represent the area of ​​the sample object in the first sample image in the form of a mask (MASK), and the embodiments of the present disclosure are not limited to the form of the position information.

[0093] In some embodiments, the sample object information may include an object type of the sample object.

[0094] In some embodiments, the sample object information may include position information of the sample object in the first sample image.

[0095] In some embodiments, the sample object information may include an object type of the sample object and position information of the sample object in the first sample image.

[0096] S103: Train the preset model according to the first sample set to obtain a target model.

[0097] The target model may be used to perform image segmentation on the first target image to obtain target object information corresponding to the first target image, wherein the target object information includes the object type of the target object and / or the position information of the target object in the first target image.

[0098] In some embodiments, the target object information may include an object type of the target object.

[0099] In some other embodiments, the target object information may include position information of the target object in the first target image.

[0100] In some other embodiments, the target object information may include the object type of the target object and the position information of the target object in the first target image.

[0101] The preset model can be any machine learning model, for example, a neural network model that can segment images.

[0102] The target object may be any object, animal, plant, portrait, human body part, animal or plant part, animal or plant species, etc., and the embodiments of the present disclosure do not limit this. The position information may be used to indicate the area of ​​the target object in the first target image. For example, the position information may be in the form of a mask (MASK) to indicate the area of ​​the target object in the first target image, and the embodiments of the present disclosure do not limit the form of the position information.

[0103] In some embodiments, based on the above-mentioned first sample set, the preset model can be trained through automated model search, tuning training, pruning and quantization, etc. to generate a target model.

[0104] Optionally, the generated target model can be sent to an automated testing unit for testing. The entire training process can be low-code, and the user only needs to select some key options.

[0105] The above method is used to obtain first data, generate a first sample set based on the first data, and train a preset model based on the first sample set to obtain a target model. The first data includes first text data; the first sample set includes a first sample image and a first sample label corresponding to the first sample image, the first sample label is used to annotate sample object information corresponding to the first sample image, and the target model can be used to perform image segmentation on the first target image to obtain target object information corresponding to the first target image, the target object information includes the object type of the target object and / or the location information of the target object in the first target image. In this way, a sample set for model training can be efficiently obtained, which improves the efficiency of model generation.

[0106] Figure 2 is based on Figure 1 The embodiment shown is a flow chart of step S102. Figure 2 As shown, the step S102 may include the following sub-steps:

[0107] S1021. Acquire second text data according to the first text data.

[0108] In some embodiments, the first text data may be used as the second text data, which is equivalent to directly generating the first sample set according to the first text data.

[0109] In other embodiments, the prompt text expansion may be performed on the first text data to obtain the second text data.

[0110] In some other embodiments, the second text data may be obtained by combining the text data obtained by performing prompt text expansion on the first text data with the first text data.

[0111] The above-mentioned prompt text expansion can also be called prompt expansion, which can provide a more detailed text description of the scene and features of the first text data. For example, the first text is "dog", and the second text data obtained after the prompt text expansion can be "there is a pure white dog" or "there are 2 puppies playing on the mahjong table indoors, one is a husky and the other is a golden retriever." Optionally, the prompt text expansion of the first text data can be performed through a pre-generated expansion model.

[0112] S1022: Generate a first sample set according to the second text data.

[0113] The first sample set may include a first sample image and a first sample label corresponding to the first sample image.

[0114] In the embodiments of the present disclosure, there may be multiple ways to generate the first sample set according to the second text data.

[0115] Figure 3 is based on Figure 2 The embodiment shown is a flow chart of step S1022. Figure 3 As shown, the step S1022 may include the following sub-steps:

[0116] S10221. Generate a first sample image according to second text data.

[0117] There are many ways to generate the first sample image according to the second text data, for example:

[0118] In one implementation, a first sample image that meets a first search condition can be obtained by searching a pre-generated first data set according to the second text data; wherein the first data set includes a plurality of sample images, and the first search condition can be used to characterize that the similarity between a first image feature of the first sample image and a text feature of the second text data is greater than or equal to a first preset similarity threshold. It should be noted that the first data set may include pre-generated labeled sample images and / or unlabeled sample images.

[0119] For example, a search may be performed in the pre-generated first data set according to the second text data to find the first sample image with the matching content. It should be noted that the second text data may also be the above-mentioned first text data.

[0120] In one implementation, the second text data may be input into a pre-generated first text image generation model to obtain a first sample image output by the first text image generation model.

[0121] By way of example, the first text image generation model may be a model such as DALL E, stable diffusion, DreamBooth, or gen1.

[0122] Optionally, the first data may also include unlabeled first image data. The first sample image may be generated in the following manner:

[0123] In one implementation, a first sample image that meets a second search condition can be retrieved from a pre-generated first data set based on the second text data and the first image data; wherein the first data set includes a plurality of sample images, and the second search condition is used to characterize that the similarity between a first image feature of the first sample image and a text feature of the second text data is greater than or equal to a second preset similarity threshold, or that the similarity between a first image feature of the first sample image and a second image feature of the first image data is greater than or equal to a third preset similarity threshold.

[0124] For example, a search may be performed in the pre-generated first data set according to the second text data or the first image data to find the first sample image with the matching content. It should also be noted that the second text data may also be the above-mentioned first text data.

[0125] In one implementation, the second text data and the first image data may be input into a pre-generated diffusion model (Diffusion Engine) to obtain a first sample image output by the diffusion model.

[0126] For example, more sample images having the same object type as the second text data or the first image data may be generated through the diffusion model.

[0127] Optionally, the first sample image may be generated by any one or more of the above methods.

[0128] S10222. Label the first sample image to obtain a first sample label corresponding to the first sample image.

[0129] The first sample label is used to mark the sample object information corresponding to the first sample image.

[0130] Optionally, the sample object information may include an object type of the sample object and / or position information of the sample object in the first sample image;

[0131] The first sample image may be manually marked or automatically marked. There are many ways to mark the first sample image, for example:

[0132] In some embodiments, the first sample image may be labeled using a pre-generated first model to obtain a first sample label corresponding to the first sample image.

[0133] The first model is a pre-trained vertical classification model, or the first model is a model obtained based on small sample segmentation training.

[0134] For example, based on the existing first model, the acquired first sample image can be labeled to obtain a first sample label corresponding to the first sample image. The first model can be a SOTA (State Of The Arts) model, which can be an accumulated vertical classification model or a multimodal segmentation model.

[0135] For another example, the first data also includes labeled second image data, which may be a small amount of pair data, and the first sample image may be labeled by small sample segmentation / Painter based on the second image data. That is, the Painter model obtained based on small sample segmentation is used as the first model, and the first sample image is labeled according to the first model to obtain the first sample label corresponding to the first sample image.

[0136] In some embodiments, a search is performed in a pre-generated first data set based on the first sample image to obtain labeled sample images with matching content.

[0137] For example, based on the first sample image, a sample image with a label that meets the fourth search condition can be retrieved in the pre-generated first data set, and the label of the sample image is used as the first sample label of the first sample image. The first data set includes multiple sample images, and the fourth search condition is used to characterize that the similarity between the first image feature of the first sample image and the image feature of the retrieved sample image is greater than or equal to a preset similarity threshold.

[0138] Figure 4 is based on Figure 2 Another flow chart of step S1022 is shown in the embodiment shown. Figure 4 As shown, the step S1022 may include the following sub-steps:

[0139] S10223. Input the second text data into a pre-generated second text image generation model to obtain a first sample image output by the second text image generation model and a first sample label corresponding to the first sample image.

[0140] The second text image generation model may be a modified text-to-image model, and the second text image generation model may output the first sample label while generating the first sample image. For example, the second text image generation model may be similar to a Grounded Diffusion model.

[0141] Figure 5 is based on Figure 2 Another flow chart of step S1022 is shown in the embodiment shown. Figure 5 As shown, the first data includes first text data and unlabeled first image data, and the step S1022 may include the following sub-steps:

[0142] S10224. Search in a pre-generated second data set according to the second text data and / or the first image data to obtain a first sample image that meets a third search condition and a first sample label corresponding to the first sample image.

[0143] The second data set includes a plurality of labeled sample images, and the third retrieval condition is used to characterize that the similarity between the first image feature of the first sample image and the image feature of the retrieved sample image is greater than or equal to a fourth preset similarity threshold.

[0144] For example, the label of the sample image in the second data set may be used to annotate the sample object information corresponding to the sample image, and the sample object information may also include the object type of the sample object and / or the position information of the sample object in the sample image.

[0145] Figure 6 is based on Figure 1 The embodiment shown is a flowchart of another step S1022. Figure 6 As shown, the first data includes first text data and second image data with a label, and the step S1022 may include the following sub-steps:

[0146] S10225. Generate multiple first sample images based on the control network according to the second image data, the second text data and the preset sample label.

[0147] Among them, the multiple first sample images correspond to the same preset sample label.

[0148] For example, based on the above-mentioned second image data (such as pair data) and the second text data, the category and shape of the object in the generated image can be controlled by a control network (controlnet), so that a large number of first sample images with the same preset sample labels (such as the same mask) can be obtained.

[0149] Thus, based on the above Figure 3 , Figure 4 , Figure 5 , Figure 6 Any one or more optional implementations of the illustrated embodiments can generate a first sample image and a first sample label corresponding to the first sample image based on the second text data, thereby obtaining a first sample set, thereby efficiently and flexibly acquiring a large number of samples.

[0150] In some embodiments of the present disclosure, the trained target model may be tested to obtain a model test index of the target model.

[0151] For example, the target model can be tested based on the automated testing module to obtain model testing indicators.

[0152] The model test index can be used to evaluate the performance and effect of the target model in detail and obtain a detailed evaluation report.

[0153] For example, testing can be performed based on a mobile terminal to obtain model test indicators of different processors on the mobile terminal, and the processors may include CPU, GPU, COREML, NPU, APU, DSP, etc.

[0154] For another example, testing can be performed based on a server to obtain model test indicators on different processors on the server. The processors may include T4, P4, A10 and other models. The model test indicators may include, for example, model time consumption, QPS (Queries-Per-Second), video memory occupancy, etc.

[0155] In some embodiments, the model testing index may include indicators for evaluating overall accuracy, edge accuracy, effect stability, effect consistency, and the like.

[0156] In some embodiments of the present disclosure, after the target model is generated, it can also be actively updated to achieve learning optimization.

[0157] Figure 7 FIG. 1 is a flow chart of a model updating method according to an embodiment of the present disclosure. Figure 7 As shown, the method may include:

[0158] S701. Obtain second data.

[0159] The second data is image data with inaccurate model segmentation obtained during testing or application of the trained target model.

[0160] S702: Generate a second sample set according to second data.

[0161] The second sample set includes a second sample image and a second sample label corresponding to the second sample image, and the second sample label is used to mark the corresponding sample object information in the second sample image.

[0162] It should be noted that the specific method of generating the second sample set according to the second data can refer to Figure 1 The manner of generating the first sample set according to the first data in the illustrated embodiment will not be described in detail here.

[0163] S703: Iteratively train the target model according to the second sample set to obtain an updated target model.

[0164] It should be noted that the second data may also be called difficult data, and the second sample set may be a difficult sample set generated based on the difficult data. By iteratively training the model with the difficult sample set, a closed loop may be formed to further improve the accuracy of the updated target model.

[0165] Figure 8 FIG. 1 is a flowchart of an image processing method according to an embodiment of the present disclosure. Figure 8 As shown, the method may include:

[0166] S801: Acquire a first target image to be processed.

[0167] S802: Input the first target image into a pre-generated target model, perform image segmentation on the first target image, and obtain target object information corresponding to the first target image.

[0168] The target object information includes the object type of the target object and / or the position information of the target object in the first target image.

[0169] In some embodiments, the target model may be a model generated by any optional method of the model generation method in the aforementioned embodiments of the present disclosure.

[0170] For example, the target model is a model obtained by training a preset model according to a first sample set, the first sample set is a sample set generated according to first data, the first data includes first text data, the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to annotate sample object information corresponding to the first sample image. Optionally, the sample object information includes an object type of the sample object and / or position information of the sample object in the first sample image.

[0171] In this way, the first target image can be segmented based on the target model generated by the above model generation method to obtain target object information corresponding to the first target image.

[0172] Fig. 9 is a block diagram of a model generation device 1100 according to an embodiment of the present disclosure, such as Fig. 9 As shown, the device 1100 may include:

[0173] A first acquisition module 1101 is used to acquire first data; wherein the first data includes first text data;

[0174] A sample generation module 1102 is configured to generate a first sample set according to the first data; wherein the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to annotate sample object information corresponding to the first sample image;

[0175] The model generation module 1103 is used to train the preset model according to the first sample set to obtain a target model; wherein the target model is used to perform image segmentation on the first target image to obtain target object information corresponding to the first target image, and the target object information includes the object type of the target object and / or the position information of the target object in the first target image.

[0176] In some embodiments, the sample generation module 1102 is used to obtain second text data according to first text data; and generate the first sample set according to the second text data.

[0177] In some embodiments, the sample generation module 1102 is used to perform prompt text expansion on the first text data to obtain second text data; or, use the first text data as the second text data.

[0178] In some embodiments, the sample generation module 1102 is used to generate a first sample image based on the second text data; the first sample image is labeled by a pre-generated first model to obtain a first sample label corresponding to the first sample image; wherein the first model is a pre-trained vertical classification model, or the first model is a model obtained based on small sample segmentation training.

[0179] In some embodiments, the sample generation module 1102 is used to search in a pre-generated first data set according to the second text data to obtain the first sample image that meets a first search condition; wherein the first data set includes a plurality of sample images, and the first search condition is used to characterize that the similarity between a first image feature of the first sample image and a text feature of the second text data is greater than or equal to a first preset similarity threshold; or,

[0180] The second text data is input into a pre-generated first text image generation model to obtain a first sample image output by the first text image generation model.

[0181] In some embodiments, the first data also includes first image data; the sample generation module 1102 is used to search in a pre-generated first data set according to the second text data and / or the first image data to obtain the first sample image that meets a second search condition; wherein the first data set includes a plurality of sample images, and the second search condition is used to characterize that the similarity between the first image feature of the first sample image and the text feature of the second text data is greater than or equal to a second preset similarity threshold, or that the similarity between the first image feature of the first sample image and the second image feature of the first image data is greater than or equal to a third preset similarity threshold; or,

[0182] The second text data and the first image data are input into a pre-generated diffusion model to obtain a first sample image output by the diffusion model.

[0183] In some embodiments, the sample generation module 1102 is used to input the second text data into a pre-generated second text image generation model to obtain a first sample image output by the second text image generation model and a first sample label corresponding to the first sample image.

[0184] In some embodiments, the first data also includes unlabeled first image data; the sample generation module 1102 is used to search in a pre-generated second data set according to the second text data and / or the first image data, to obtain the first sample image that meets a third retrieval condition and the first sample label corresponding to the first sample image; wherein the second data set includes multiple labeled sample images, and the third retrieval condition is used to characterize that the similarity between the first image feature of the first sample image and the image feature of the retrieved sample image is greater than or equal to a fourth preset similarity threshold.

[0185] In some embodiments, the first data also includes labeled second image data; the sample generation module 1102 is used to generate multiple first sample images based on the control network according to the second image data, the second text data and the preset sample label; wherein the multiple first sample images correspond to the same preset sample label.

[0186] In some embodiments, the apparatus 1100 further includes:

[0187] The model testing module 1104 is used to test the trained target model and obtain the model testing index of the target model.

[0188] In some embodiments, the apparatus 1100 further includes:

[0189] The model updating module 1105 is used to obtain second data; the second data is image data with inaccurate model segmentation obtained during testing or application of the trained target model; a second sample set is generated based on the second data; wherein the second sample set includes a second sample image and a second sample label corresponding to the second sample image, and the second sample label is used to annotate the corresponding sample object information in the second sample image; the target model is iteratively trained based on the second sample set to obtain an updated target model.

[0190] Fig.10 is a block diagram of an image processing device 1200 according to an embodiment of the present disclosure, such as Fig.10 As shown, the apparatus 1200 may include:

[0191] The second acquisition module 1201 is used to acquire a first target image to be processed;

[0192] An image processing module 1202 is used to input the first target image into a pre-generated target model, perform image segmentation on the first target image, and obtain target object information corresponding to the first target image, wherein the target object information includes an object type of the target object and / or position information of the target object in the first target image;

[0193] Among them, the target model is a model obtained by training a preset model according to a first sample set, the first sample set is a sample set generated according to first data, the first data includes first text data, the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to mark the sample object information corresponding to the first sample image.

[0194] In some embodiments, the target model is a model generated according to the method described in the first aspect of the present disclosure.

[0195] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0196] Reference below Fig.11, which shows a schematic diagram of the structure of an electronic device 2000 (e.g., a terminal device or a server) suitable for implementing the embodiment of the present disclosure. The terminal device in the embodiment of the present disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (e.g., vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. The server in the embodiment of the present disclosure may include, but is not limited to, local servers, cloud servers, single servers, distributed servers, etc. Fig.11 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0197] like Fig.11 As shown, the electronic device 2000 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 2001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 2002 or a program loaded from a storage device 2008 into a random access memory (RAM) 2003. In the RAM 2003, various programs and data required for the operation of the electronic device 2000 are also stored. The processing device 2001, the ROM 2002, and the RAM 2003 are connected to each other via a bus 2004. An input / output (I / O) interface 2005 is also connected to the bus 2004.

[0198] Typically, the following devices may be connected to the input / output interface 2005: an input device 2006 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 2007 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 2008 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 2009. The communication device 2009 may allow the electronic device 2000 to communicate with other devices wirelessly or by wire to exchange data. Although Fig.11 The electronic device 2000 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0199] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 2009, or installed from the storage device 2008, or installed from the ROM 2002. When the computer program is executed by the processing device 2001, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0200] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. Computer-readable storage media may include, but are not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer 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 disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0201] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0202] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0203] The above-mentioned computer-readable medium carries one or more programs. When the above-mentioned one or more programs are executed by the electronic device, the electronic device: obtains first data; wherein the first data includes first text data; generates a first sample set based on the first data; wherein the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to annotate the sample object information corresponding to the first sample image; trains a preset model according to the first sample set to obtain a target model; wherein the target model is used to perform image segmentation on the first target image to obtain target object information corresponding to the first target image, and the target object information includes the object type of the target object and / or the position information of the target object in the first target image.

[0204] Alternatively, the computer-readable medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device: obtains a first target image to be processed; inputs the first target image into a pre-generated target model, performs image segmentation on the first target image, and obtains target object information corresponding to the first target image, wherein the target object information includes the object type of the target object and / or the location information of the target object in the first target image; wherein the target model is a model obtained by training a preset model according to a first sample set, the first sample set is a sample set generated according to first data, the first data includes first text data, the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to annotate the sample object information corresponding to the first sample image.

[0205] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0206] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0207] The modules involved in the embodiments described in the present disclosure may be implemented by software or hardware. The name of a module does not limit the module itself in some cases. For example, the first acquisition module may also be described as a "module for acquiring first data".

[0208] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0209] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer 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 disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0210] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0211] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0212] Although the subject matter has been described in language specific to structural features and / or method logic actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims. Regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

Claims

1. A model generation method, characterized in that: The method comprises: Acquire first data; wherein the first data includes first text data; Generate a first sample set according to the first data; wherein the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to mark sample object information corresponding to the first sample image; The preset model is trained according to the first sample set to obtain a target model; wherein the target model is used to perform image segmentation on the first target image to obtain target object information corresponding to the first target image, and the target object information includes the object type of the target object and / or the location information of the target object in the first target image.

2. The method according to claim 1, characterized in that Generating a first sample set according to the first data includes: Acquire second text data according to the first text data; The first sample set is generated according to the second text data.

3. The method according to claim 2, characterized in that The acquiring of the second text data according to the first text data comprises: Performing prompt text expansion on the first text data to obtain second text data; or, The first text data is used as the second text data.

4. The method according to claim 2, characterized in that: Generating the first sample set according to the second text data includes: generating a first sample image according to the second text data; The first sample image is labeled by using a pre-generated first model to obtain a first sample label corresponding to the first sample image; wherein the first model is a pre-trained vertical classification model, or the first model is a model obtained based on small sample segmentation training.

5. The method according to claim 4, characterized in that Generating a first sample image according to the second text data comprises: According to the second text data, the first sample image that meets the first search condition is searched in the pre-generated first data set; wherein the first data set includes a plurality of sample images, and the first search condition is used to characterize that the similarity between the first image feature of the first sample image and the text feature of the second text data is greater than or equal to a first preset similarity threshold; or The second text data is input into a pre-generated first text image generation model to obtain a first sample image output by the first text image generation model.

6. The method according to claim 4, characterized in that The first data also includes first image data; and generating a first sample image according to the second text data includes: According to the second text data and / or the first image data, a search is performed in a pre-generated first data set to obtain the first sample image that meets a second search condition; wherein the first data set includes a plurality of sample images, and the second search condition is used to characterize that the similarity between the first image feature of the first sample image and the text feature of the second text data is greater than or equal to a second preset similarity threshold, or that the similarity between the first image feature of the first sample image and the second image feature of the first image data is greater than or equal to a third preset similarity threshold; or, The second text data and the first image data are input into a pre-generated diffusion model to obtain a first sample image output by the diffusion model.

7. The method according to claim 2, characterized in that Generating the first sample set according to the second text data includes: The second text data is input into a pre-generated second text image generation model to obtain a first sample image output by the second text image generation model and a first sample label corresponding to the first sample image.

8. The method according to claim 2, characterized in that: The first data also includes unlabeled first image data; and generating the first sample set according to the second text data includes: According to the second text data and / or the first image data, a search is performed in a pre-generated second data set to obtain the first sample image that meets a third search condition and a first sample label corresponding to the first sample image; wherein the second data set includes a plurality of labeled sample images, and the third search condition is used to characterize that the similarity between the first image feature of the first sample image and the image feature of the retrieved sample image is greater than or equal to a fourth preset similarity threshold.

9. The method according to claim 2, characterized in that: The first data also includes second image data with a label; and generating the first sample set according to the second text data includes: According to the second image data, the second text data and the preset sample label, a plurality of first sample images are generated based on a control network; wherein the plurality of first sample images correspond to the same preset sample label.

10. The method according to any one of claims 1 to 9, characterized in that The method further comprises: The trained target model is tested to obtain a model test index of the target model.

11. The method according to any one of claims 1 to 9, characterized in that The method further comprises: Acquire second data; the second data is image data with inaccurate model segmentation obtained by testing or applying the trained target model; Generate a second sample set according to the second data; wherein the second sample set includes a second sample image and a second sample label corresponding to the second sample image, and the second sample label is used to mark the corresponding sample object information in the second sample image; The target model is iteratively trained according to the second sample set to obtain an updated target model.

12. An image processing method, characterized in that: The method comprises: Acquire a first target image to be processed; Inputting the first target image into a pre-generated target model, performing image segmentation on the first target image, and obtaining target object information corresponding to the first target image, wherein the target object information includes an object type of the target object and / or position information of the target object in the first target image; Among them, the target model is a model obtained by training a preset model according to a first sample set, the first sample set is a sample set generated according to first data, the first data includes first text data, the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to mark the sample object information corresponding to the first sample image.

13. A model generation device, characterized in that: The device comprises: A first acquisition module, used to acquire first data; wherein the first data includes first text data; A sample generation module, configured to generate a first sample set according to the first data; wherein the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to mark sample object information corresponding to the first sample image; A model generation module is used to train a preset model according to the first sample set to obtain a target model; wherein the target model is used to perform image segmentation on a first target image to obtain target object information corresponding to the first target image, and the target object information includes the object type of the target object and / or the location information of the target object in the first target image.

14. An image processing device, characterized in that: The device comprises: A second acquisition module, used for acquiring a first target image to be processed; an image processing module, configured to input the first target image into a pre-generated target model, perform image segmentation on the first target image, and obtain target object information corresponding to the first target image, wherein the target object information includes an object type of the target object and / or position information of the target object in the first target image; Among them, the target model is a model obtained by training a preset model according to a first sample set, the first sample set is a sample set generated according to first data, the first data includes first text data, the first sample set includes a first sample image and a first sample label corresponding to the first sample image, and the first sample label is used to mark the sample object information corresponding to the first sample image.

15. A computer readable medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processing device, the steps of the method described in any one of claims 1 to 11 are implemented; or, when the computer program is executed by a processing device, the steps of the method described in claim 12 are implemented.

16. An electronic device, characterized in that: include: a storage device having a computer program stored thereon; A processing device, used to execute the computer program in the storage device to implement the steps of the method according to any one of claims 1 to 11, or to implement the steps of the method according to claim 12.

Images