Intelligent image processing method, readable storage medium, computer program product

Abstract

This application discloses an intelligent image processing method, a readable storage medium, and a computer program product. The method includes: receiving product reviews submitted by a user for a target product, and determining a target image to be processed from the review; the target image includes image content generated by photographing the outer packaging of the target product, and the image content includes a text region containing text information on the outer packaging; generating a first image processing scheme for the text region, the first image processing scheme including a model identifier and first model parameters of a first super-resolution model, and a first image enhancement method and a corresponding first intensity parameter; calling the first super-resolution model so that the model performs super-resolution reconstruction of the text region according to the first model parameters, and outputs a super-resolution processed image; performing image enhancement processing on the super-resolution processed image according to the first image enhancement method and the first intensity parameter, so as to publish product review information based on the intelligently image-processed target image.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to an intelligent image processing method, a computer-readable storage medium, an electronic device, and a computer program product. Background Technology

[0002] In online shopping scenarios, consumers can submit online orders through applications provided by the product information service system. After the order is fulfilled, they can provide product review services, whereby users can publish reviews related to the ordered products in the form of pictures and text.

[0003] Analysis of published reviews revealed that most user-posted review images suffered from issues such as blurry images and distorted colors, resulting in low reference value for the reviews and potentially influencing the purchasing decisions of other users who viewed them.

[0004] Improving the image quality of user-posted images has become a technical problem that needs to be solved by those skilled in the art. Summary of the Invention

[0005] This application provides an intelligent image processing method and apparatus, a computer-readable storage medium, an electronic device, and a computer program product, which helps to improve the image quality of user-posted product review images, thereby increasing the reference value of user-posted product review content.

[0006] This application provides the following solution: A smart image processing method, applied on a server, the method comprising: The system receives product reviews submitted by users for a target product and determines a target image to be processed from the product review content. The target image includes image content generated by taking a picture of the outer packaging of the target product, and the image content includes the text area where the text information on the outer packaging is located. A first image processing scheme is generated for the text region. The first image processing scheme includes: a model identifier and first model parameters of a first super-resolution model, and a first image enhancement method and a corresponding first intensity parameter. The first super-resolution model is invoked so that the first super-resolution model performs super-resolution reconstruction of the text region of the target image according to the first model parameters and outputs the super-resolution processed image. Based on the first image enhancement method and the first intensity parameter, the super-resolution processed image is subjected to image enhancement processing so as to publish product evaluation information based on the target image after intelligent image processing.

[0007] Wherein, if the target product is related to the food category, the target image also includes product image content related to the food, and the method further includes: Identify the sub-category of the target product under the food category, obtain image processing knowledge information related to the sub-category and related to enhancing the appetite expression of the image, and generate a second image processing scheme for the image region where the product image is located based on the knowledge information. The second image processing scheme includes a second image enhancement method and a corresponding second intensity parameter. After obtaining the super-resolution processed image, image enhancement processing is performed on the image region according to the second image enhancement method and the second intensity parameter.

[0008] The method further includes: The target image is subjected to image quality assessment, and if the assessment result indicates that the target image needs to be super-resolution processed, a third image processing scheme for the target image is generated. The third image processing scheme includes the model identifier of the second super-resolution model and the second model parameters. The second super-resolution model is invoked so that the second super-resolution model performs super-resolution reconstruction on the super-resolution processed image output by the first super-resolution model according to the second model parameters, and the output of the second super-resolution model is used as the super-resolution processed image for image enhancement processing.

[0009] The method further includes: If the evaluation result indicates that super-resolution processing of the target image is not required, the image region is compressed with the goal of maintaining image integrity, and the compressed target image is used as the input of the first super-resolution model for super-resolution processing.

[0010] The step of performing image quality assessment on the target image includes: Obtain the first evaluation dimension information and the corresponding first weight information associated with the text region, and the second evaluation dimension information and the corresponding second weight information associated with the subdivided category; The text region is evaluated for image quality based on the first evaluation dimension information and the first weight information to obtain a first evaluation score. The image region is evaluated for image quality based on the second evaluation dimension information and the second weight information to obtain a second evaluation score. The evaluation result of the target image is obtained based on the first evaluation score and the second evaluation score.

[0011] The method further includes: The text region of the target image is sharpened, and the sharpened target image is used as the input of the first super-resolution model for super-resolution processing.

[0012] The first model parameter includes image block information, which is used to indicate that the target image is not cut into image blocks.

[0013] The second model parameters include image block information, which indicates that the target image is cut into a preset number of image blocks so that the second super-resolution model can perform super-resolution reconstruction on the image blocks respectively. The method further includes: After performing super-resolution processing and image enhancement processing on the image blocks respectively, image block fusion is performed to obtain the target image after intelligent image processing.

[0014] If the target image does not include the text region, the method further includes: The system determines that the target product is related to the food category, and that the target image includes product image content related to the food. It also identifies the sub-category of the target product under the food category and performs image quality assessment on the target image. Based on the image processing knowledge information related to enhancing the appetite expression of the image associated with the subdivided category and the evaluation results of the image quality assessment, a fourth image processing scheme is generated for the target image. The fourth image processing scheme includes: a third image enhancement method and a corresponding third intensity parameter, and a model identifier and third model parameters of a third super-resolution model determined when the evaluation result indicates that the target image needs to be super-resolution processed. The third super-resolution model is invoked so that the third super-resolution model can perform super-resolution reconstruction of the target image according to the parameters of the third model and output the super-resolution processed image. Based on the third image enhancement method and the third intensity parameter, the super-resolution processed image is subjected to image enhancement processing so as to publish product evaluation information based on the target image after intelligent image processing.

[0015] The method further includes: If a user submits an image adjustment request for the target image, the system determines the image adjustment method and parameters corresponding to the request, and adjusts the target image accordingly, so as to publish product review information based on the adjusted target image.

[0016] A smart image processing method, applied on a server, the method comprising: The system receives product reviews submitted by users for a target product and determines the target image to be processed from the product reviews; the target image includes an image generated by taking a picture of the outer packaging of the target product. If the target product is related to the food category, and the target image includes product image content related to the food, then the sub-category of the target product under the food category is identified, and image processing knowledge information related to the sub-category and improving the appetite-enhancing effect of the image is obtained. An image optimization scheme is generated based on the knowledge information, and intelligent image processing is performed on the target image so as to publish product evaluation information based on the intelligently processed target image.

[0017] A smart image processing method, applied to a client, the method comprising: The system receives product reviews submitted by users for a target product and sends them to the server. The server then determines the target image to be processed from the product review content. The target image includes an image of the outer packaging of the target product, and the image content includes a text area containing text information on the outer packaging. A first image processing scheme is generated for the text area, and a first super-resolution model is called to perform super-resolution reconstruction of the text area according to the first image processing scheme. Image enhancement processing is then performed on the super-resolution processed image so that product review information can be published based on the intelligently image-processed target image.

[0018] The method further includes: Receive product reviews submitted by users for the target product and provide a first operation option for initiating intelligent image processing; After obtaining an image processing request for the target image in the product review content through the first operation option, the product review content is sent to the server.

[0019] The method further includes: Provide a second operation option for submitting image adjustment requests; The second operation option obtains the image adjustment request submitted by the user for the target image and sends it to the server so that the server can determine the image adjustment method and adjustment parameters corresponding to the image adjustment request, perform image adjustment on the target image, and obtain the adjusted target image.

[0020] The second operation option consists of tabs provided for different image adjustment methods. The step of obtaining the image adjustment request submitted by the user for the target image through the second operation option includes: when the target tab is selected, determining the image adjustment method represented by the target tab as the image adjustment request submitted by the user.

[0021] The second operation option is an information input box. The step of obtaining the image adjustment request submitted by the user for the target image through the second operation option includes: obtaining the image adjustment request submitted by the user through the information input box.

[0022] A smart image processing method is applied to the application client of a commodity information service system, the method comprising: The system receives user-generated content (UGC) posted by a user for a target product and sends it to the server. The server then determines the target image to be processed from the UGC, where the target image includes user experience information related to the target product. A first image processing scheme is generated for the text region containing the user experience information. Based on the first image processing scheme, a first super-resolution model is invoked to perform super-resolution reconstruction on the text region. Image enhancement processing is then performed on the super-resolution image to obtain a target image after intelligent image processing. User-generated content is then posted based on this intelligent image-processed target image.

[0023] A smart image processing device, applied on a server side, the device comprising: The target image determination unit is used to receive product evaluation content submitted by a user for a target product, and determine the target image to be processed from the product evaluation content; the target image includes image content generated by taking a picture of the outer packaging of the target product, and the image content includes the text area where the text information on the outer packaging is located; An image processing scheme generation unit is used to generate a first image processing scheme for the text region. The first image processing scheme includes: a model identifier and first model parameters of a first super-resolution model, and a first image enhancement method and a corresponding first intensity parameter. The super-resolution model invocation unit is used to invoke the first super-resolution model so that the first super-resolution model can perform super-resolution reconstruction of the text region of the target image according to the first model parameters and output the super-resolution processed image. An image enhancement processing unit is used to perform image enhancement processing on the super-resolution processed image according to the first image enhancement method and the first intensity parameter, so as to publish product evaluation information based on the target image after intelligent image processing.

[0024] A smart image processing device, applied on a server side, the device comprising: The target image determination unit is used to receive product evaluation content submitted by a user for a target product, and determine the target image to be processed from the product evaluation content; the target image includes image content generated by taking a picture of the outer packaging of the target product; The sub-category identification unit is used to identify the sub-category of the target product under the food category, where the target product is related to the food category and the target image includes product image content related to the food, and to obtain image processing knowledge information related to the sub-category and related to improving the appetite-enhancing effect of the image. The image optimization scheme generation unit is used to generate an image optimization scheme based on the knowledge information and perform intelligent image processing on the target image so as to publish product evaluation information based on the intelligently processed target image.

[0025] A smart image processing device, applied to a client, the device comprising: A product review content sending unit is used to receive product review content submitted by users for a target product and send it to the server so that the server can determine the target image to be processed from the product review content. The target image includes image content generated by taking a picture of the outer packaging of the target product, and the image content includes a text area containing text information on the outer packaging. A first image processing scheme is generated for the text area, and a first super-resolution model is called according to the first image processing scheme to perform super-resolution reconstruction of the text area, and image enhancement processing is performed on the super-resolution processed image so as to publish product review information based on the target image after intelligent image processing.

[0026] A smart image processing device is used in the application client of a commodity information service system. The device includes: The product review content sending unit is used to receive user-generated content posted by users for a target product and send it to the server so that the server can determine the target image to be processed from the user-generated content. The target image includes user experience information related to the target product. A first image processing scheme is generated for the text area where the user experience information is located. The server calls a first super-resolution model to perform super-resolution reconstruction on the text area according to the first image processing scheme and performs image enhancement processing on the super-resolution processed image to obtain the target image after intelligent image processing, so that user-generated content can be published based on the target image after intelligent image processing.

[0027] A computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the steps of any of the preceding methods.

[0028] An electronic device, comprising: One or more processors; and A memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform the steps of any of the preceding methods.

[0029] A computer program product includes a computer program / computer executable instructions that, when executed by a processor in an electronic device, implement the steps of any of the preceding methods.

[0030] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application embodiment can determine the target image to be processed from the product review content submitted by users for the target product, and provide image optimization processing for the target image before the review is published to ensure the high-quality publication of the product review image.

[0031] Specifically, when it is determined that the image content generated from the outer packaging of the target product includes text information on the outer packaging, the server can generate a first image processing scheme for the text area where the text information is located.

[0032] The image processing procedure of this application can be implemented in two stages. First, the text region can be super-resolution processed using a super-resolution model to optimize image quality by increasing the number of pixels. Second, the text region can be further enhanced to improve image quality by increasing pixel quality. Therefore, the first image processing scheme generated by the server can include: a model identifier and first model parameters of the first super-resolution model used for super-resolution processing; and a first image enhancement method and corresponding first intensity parameters used for image enhancement processing.

[0033] The server performs super-resolution model calls and image enhancement processing according to the first image processing scheme to obtain the target image after intelligent image processing, which can be used when publishing product evaluation information.

[0034] This approach helps improve the image quality of target images, thereby increasing the reference value of user-submitted product reviews. For users submitting product reviews, it helps identify their reviews as high-quality and give them special markings, improving the user experience. For users viewing product reviews, high-quality reviews help enhance the browsing experience and facilitate quick purchasing decisions, thus increasing user shopping efficiency and the store's product conversion rate.

[0035] Of course, any product implementing this application does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a schematic diagram of the intelligent image processing system provided in the embodiments of this application; Figure 2 This is a schematic diagram of a page provided in an embodiment of this application; Figure 3 This is a flowchart of an intelligent image processing method provided in an embodiment of this application; Figure 4 This is a flowchart of another intelligent image processing method provided in the embodiments of this application; Figure 5 This is a schematic diagram of an intelligent image processing device provided in an embodiment of this application; Figure 6 This is a schematic diagram of another intelligent image processing device provided in an embodiment of this application; Figure 7 This is a schematic diagram of the electronic device provided in the embodiments of this application. Detailed Implementation

[0038] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0039] Analysis of published product reviews revealed that most users do not pay much attention to the quality of the product review images they post. Images taken casually often suffer from cluttered backgrounds, blurry images, and color distortion. Therefore, this application provides an optimization scheme for product review images. Posting product reviews based on optimized images helps improve the quality of user-generated image and text reviews.

[0040] Specifically, from a system architecture perspective, the intelligent image processing system of this application embodiment can be as follows: Figure 1 As shown, this includes both the client and the server.

[0041] As an example, this application can optimize review images without the user's awareness. Specifically, the client can obtain product review content submitted by the user for the target product, and send the product review content to the server when the user performs a review content publishing operation, such as clicking the publish button. Correspondingly, the server can determine the target image to be processed from the product review content and optimize the target image. Subsequently, product review information can be published based on the image-processed target image. In other words, this application can perform image optimization processing by the server without the user's awareness without changing the existing product review publishing operation chain, without requiring additional user operation.

[0042] As another example, this application can initiate an intelligent image processing process based on the user's choice. Specifically, after receiving the product review content submitted by the user for the target product, the client can provide a first operation option to initiate intelligent image processing; after obtaining the image processing request for the target image in the product review content through the first operation option, the product review content is sent to the server deployed on the cloud server. As an example, the first operation option can be manifested as an operation button with an intelligent icon.

[0043] In practical applications, after a user submits a product review for a target product, they can click an operation button to initiate the intelligent image processing process. Correspondingly, the client can submit the product review to the server for image optimization based on the user's action, obtaining the processed target image. Subsequently, the review content publishing process can be executed directly based on the processed target image, or the processed target image can be returned to the client for display, allowing the user to view the image processing effect before deciding whether to proceed with the review content publishing process.

[0044] In this example, the client can also provide a target image selection service, whereby the user determines the target image from at least one evaluation image submitted for the target product, so that the server can perform targeted intelligent image processing on the target image selected by the user.

[0045] In one approach, if a user submits multiple review images for a target product, the client can provide a unified button to initiate the intelligent image processing procedure. The user can then select a target image from the multiple review images and click the unified button to start the image processing for that target image.

[0046] In another approach, if a user submits multiple review images for a target product, the client can provide separate action buttons for each review image. For example, it could be like this: Figure 2As shown, an operation button to start the intelligent image processing process is added at a preset position on the evaluation image. Then, when a user clicks the operation button corresponding to a certain evaluation image, the client can designate that evaluation image as the target image and start the image processing process for that target image.

[0047] Furthermore, this application embodiment can also support user-defined image processing requests. Specifically, the client can also provide a second operation option for submitting image adjustment requests. After obtaining the user's image adjustment request for the target image through the second operation option, the client sends it to the server. The server determines the image adjustment method and adjustment parameters corresponding to the image adjustment request, and adjusts the target image accordingly. The adjusted target image is then returned to the client for display. This solution enables a more flexible and personalized image processing process, meeting user-defined image adjustment needs.

[0048] The implementation process of the intelligent image processing method of this application embodiment will be explained below with specific examples. See [link to relevant documentation]. Figure 3 The flowchart shown may include: S301: The client receives product reviews submitted by users for the target product and sends them to the server.

[0049] Taking the user-initiated intelligent image processing process as an example, the client in this embodiment can provide [services] based on user operations. Figure 2 The product review page shown here allows users to edit product reviews in the text and image editing area, and upload at least one product review image via the "Upload Image" option. Product review images can take various forms, including photos, videos, and GIFs. The client can provide a button to activate intelligent image processing at a preset location on the submitted product review image.

[0050] In the intelligent image optimization scenario, if the user decides to optimize image 1, clicking the "Intelligent Optimization" button corresponding to image 1 will directly generate an image processing request for image 1 and send it to the server. The server will then determine image 1 as the target image for optimization. In this example, the "Intelligent Optimization" buttons provided for different evaluation images are the first operation options in this embodiment.

[0051] In scenarios that simultaneously support intelligent image optimization and user-defined requests, after a user clicks the "Intelligent Optimization" button corresponding to Image 1, they can be redirected to the next level page. This page provides a first operation option to initiate intelligent image processing and a second operation option to submit image adjustment requests. In this example, the "Intelligent Adjustment" tab is the first operation option in this embodiment. After the user clicks this tab, the client can generate an image processing request for Image 1 and send it to the server. The server then identifies Image 1 as the target image for image optimization.

[0052] For the second option of submitting user-defined requests, this could be represented by separate tabs for different image adjustment methods. For example... Figure 2 The tabs shown relate to image adjustment methods such as sharpness, contrast, background blur, and focus. Users can submit image adjustment requests by selecting these tabs. For example, if a user clicks the "Sharpness Enhancement" tab, the client can designate it as the target tab. Then, with the target tab selected, the image adjustment method represented by the target tab is identified as the user's submitted image adjustment request and sent to the server for targeted image optimization.

[0053] In addition, the second operation option can also be presented as an information input box, such as a text input box or a voice input box. After the user edits and submits their custom request through the information input box, the client can send the image adjustment request obtained through the information input box to the server, and the server will perform targeted image optimization.

[0054] In this embodiment, a single image processing step can be performed on a single image; alternatively, in practical applications, a single image processing step can also be performed on multiple images, meaning that the user can identify multiple target images from the submitted evaluation images and request the server to perform batch image processing. The specific configuration can be flexibly tailored to the usage requirements.

[0055] S302: The server receives product evaluation content and determines the target image to be processed from it; the target image includes the image content generated by taking a picture of the outer packaging of the target product, and the image content includes the text area where the text information on the outer packaging is located.

[0056] S303: The server generates the first image processing scheme for the text region.

[0057] Analysis of product reviews revealed that users viewing product reviews may focus on textual information within the review images, particularly key information such as ingredient lists, shelf life, and product standards. Therefore, this application can identify preset textual information from the key information that users are interested in and prioritize intelligent image processing of the text areas containing this preset textual information.

[0058] Specifically, the server identifies the target image to be processed and performs text recognition. If the target image contains preset text information, it can make intelligent decisions and generate a first image processing scheme for the text region containing the text information. That is, it determines the image optimization strategy for the text region, which can then be used to perform image optimization processing on the text region of the target image.

[0059] The embodiments of this application can perform image optimization processing in two stages. First, super-resolution processing can be performed to optimize image quality by increasing the number of pixels in the target image; second, image enhancement processing can be performed to further optimize image quality by improving the pixel quality of the target image. Correspondingly, the first image processing scheme may include: a model identifier and first model parameters of a first super-resolution model for super-resolution processing; and a first image enhancement method and a corresponding first intensity parameter for image enhancement processing.

[0060] The first super-resolution model can be a text super-resolution model. As an example, this application embodiment can train a text super-resolution model based on the capabilities of traditional models. It fully utilizes the fast response capability and high vertical domain accuracy of traditional models to achieve model inference. The traditional model can be a model built based on mathematical equations and predefined rules, mainly operating through rule mapping, with a relatively small parameter scale and characteristics such as deterministic output and strong interpretability. Alternatively, it can be trained based on the capabilities of artificial intelligence (AI) models. The AI ​​model can be a deep learning model containing massive parameters. Due to its large parameter scale, such models can store and process large amounts of information, thereby achieving higher performance on various tasks.

[0061] In the process of super-resolution model inference, the model parameters involved may include: (1) Super-resolution factor, that is, the magnification ratio of resolution, which can be reflected as 2x super-resolution, 4x super-resolution, etc. The specific value can be determined and flexibly configured according to actual usage requirements and image processing computing power. (2) Image block information, that is, the number of image blocks to be cut, which can be reflected as full image processing (i.e., no blocks), 4 blocks, 8 blocks, etc. The value can also be flexibly determined and configured. (3) Overlapping pixel value, which is mainly used for fusion processing after image block division, to eliminate the stitching traces between image blocks, avoid obvious gaps or edge breaks, and ensure that the transition of image blocks into a whole image is more natural. The value can be reflected as 32 pixels, 64 pixels, etc. The value can also be flexibly determined and configured.

[0062] As an example, the optimization strategy for text region super-resolution processing in this application can be embodied as: a 4x text super-resolution model, and without cutting image patches from the target image, i.e., performing full-image processing.

[0063] Furthermore, image enhancement methods can include noise suppression, color correction, detail enhancement, sharpness enhancement, etc., and the specific methods can be determined and configured according to actual usage requirements. As an example, the optimization strategy for text region image enhancement processing in this application can be expressed as: text region contrast +30%, where contrast is the first image enhancement method, and the 30% increase is the first intensity parameter corresponding to this method.

[0064] S304: The server calls the first super-resolution model so that the first super-resolution model can perform super-resolution reconstruction of the text region of the target image according to the parameters of the first model and output the super-resolution processed image.

[0065] After generating the first image processing scheme for the text region of the target image, the first stage of image optimization processing can be initiated, with the server calling the model based on the model identifier. Specifically, the target image and the first model parameters can be used as input to call the first super-resolution model. The model then performs super-resolution reconstruction of the text region of the target image based on the first model parameters and outputs the super-resolution processed image.

[0066] Through model super-resolution processing, pixel information that was not originally present in the target image can be inferred and generated, reconstructing more high-resolution details, effectively increasing the actual information content of the target image, and achieving the goal of improving image quality. For details regarding the model training process and the implementation process of super-resolution processing, please refer to relevant technologies; this application does not impose specific limitations on them.

[0067] To further improve the super-resolution processing effect, the text region of the target image can be sharpened first to enhance the edge contrast of the text, making the blurry and light text outlines clear and sharp. Then, the sharpened target image is used as the input of the first super-resolution model, and the model performs super-resolution reconstruction of the text region.

[0068] S305: The server performs image enhancement processing on the super-resolution processed image according to the first image enhancement method and the first intensity parameter, so as to publish product evaluation information based on the target image after intelligent image processing.

[0069] After the first super-resolution model outputs the super-resolution processed image, the second stage of image optimization processing can be started. Image enhancement processing is performed according to the first image enhancement method and its corresponding first intensity parameter, which can also be called image post-processing, to obtain the target image after intelligent image processing.

[0070] Image enhancement improves the subjective visual effect of a target image, making it clearer and more aesthetically pleasing. Understandably, this step can focus on enhancing the text area to ensure the visual quality of key information of interest to the user. For example, following the optimization strategy in the example above, enhancing the text area with a 30% contrast increase makes small text in the target image clearly readable. Optionally, in practical applications, to further highlight text information, image enhancement schemes can be configured to apply to areas outside the text area in the optimization strategy. For example, the optimization strategy in the example above can be adjusted to: 30% contrast increase for the text area, Gaussian blur for the background (σ=1.5), and removal of moiré patterns from the screenshot. In other words, in addition to enhancing the text area, image enhancement is also applied to the image background and overall moiré interference. Following this optimization strategy, small text in the target image becomes clearly readable, while also ensuring a clean background and significantly improving the credibility of the information.

[0071] S306: The client receives the target image after intelligent image processing sent by the server and displays it.

[0072] Once the server completes image optimization of the target image, it can return the intelligently processed target image to the client to demonstrate the image optimization effect to the user.

[0073] Optionally, to optimize the user experience, after the user submits an image processing request through the first operation option, the client can also obtain the image optimization progress from the server and display it synchronously. When the optimization progress indicates that the image optimization is complete, the client can obtain the processed target image returned by the server and display the effect.

[0074] In summary, this allows for image optimization of text areas containing preset text information, resulting in clearer, more visually appealing processed images that can be provided to users when making product reviews. This enhances the reference value of user-generated product reviews and improves the browsing experience for other users viewing product reviews.

[0075] In practical applications, the target image generated by a user taking a photo of the outer packaging of a target product will most likely also include image content related to the product. The image quality of the image area, especially when the target product is related to the food category, is crucial to the overall visual effect of the image. Correspondingly, the embodiments of this application can also provide the following preferred solution for intelligent image processing.

[0076] Preferred Option 1 Low-quality user-posted images may also result in unappetizing images and text. This solution addresses this by performing targeted image enhancement on the image area based on the product category of the target product when the target product is determined to be related to the food category and the target image also includes food-related product images. This will improve the visual appeal of the target image.

[0077] Taking a user's review of purchased chilled pork belly as an example, the submitted review image includes both the product label outside the packaging and the pork belly itself. During the server-side intelligent optimization of the target image, it can identify preset text information in the product label area and perform targeted image optimization on the text area according to the above scheme. Simultaneously, it can also identify the product image inside the packaging. If the target product is determined to be related to a food category based on the product image, it can further identify the sub-category of the target product within the food category, obtain image processing knowledge information related to the sub-category and enhancing the image's appeal to appetite, and generate a second image processing scheme for the image area containing the product image based on this knowledge information. The second image processing scheme may include a second image enhancement method and a corresponding second intensity parameter.

[0078] As an example, image processing knowledge can be represented as prior knowledge related to appetite expression, and the server can generate image processing solutions corresponding to specific product categories based on the prior knowledge.

[0079] As another example, image processing knowledge can be embodied in pre-configuring image enhancement methods and corresponding intensity parameters for different subcategories, with the goal of improving the appetizing expressiveness of an image. After identifying the subcategory of the target product, configuration information can be matched to determine the image enhancement methods and intensity parameters associated with the subcategory to which the target product belongs, and a second image processing scheme can be generated accordingly.

[0080] In this embodiment, the subcategories under the food category can be represented by information such as ready-to-eat meals and fresh produce. Alternatively, based on usage needs, further subcategories can be made into categories such as meat, eggs, fruits, and vegetables, or even categories such as pork belly, steak, and chicken breast. The specific granularity of the subcategories can be determined according to usage needs, and this embodiment does not limit this.

[0081] This application's embodiments can determine the configuration information associated with different subcategories of food in terms of appetite enhancement by combining the product characteristics of those subcategories. Examples are provided below.

[0082] For pork belly, the product characteristics related to enhancing appetite can be reflected in: highlighting the whiteness of the fat and the clear texture of the muscle, thus enhancing the perception of freshness and achieving the goal of enhancing the appetite of pork belly. The corresponding configuration information for pork belly sub-categories can be: fat area saturation +15% and sharpening of muscle fibers.

[0083] For steak, the product characteristics related to enhancing appetite can be reflected in: enhancing the redness and three-dimensionality of the texture. Therefore, the meat quality can be improved to enhance the appetite of steak. The corresponding configuration information for steak subcategories can be: red meat color correction and directional texture enhancement.

[0084] For chicken breast, the product characteristics related to improving appetite can be reflected in: enhancing transparency and fiber sharpness. Therefore, a healthy and low-fat approach can be adopted to enhance the appetite of chicken breast. The corresponding configuration information for chicken breast sub-categories can be: defogging treatment and high consistency fiber sharpening.

[0085] For oranges, the product characteristics related to enhancing appetite can be reflected in: bright orange color and clean background. Therefore, the freshness can be enhanced to increase the appetite of oranges. The corresponding configuration information for orange subcategories can be: enhanced orange tone and removal of background noise.

[0086] For apples, the product characteristics related to improving appetite can be reflected in: a bright red color and a moist skin. Therefore, improving the texture of the fruit can enhance the apple's appeal. The corresponding configuration information for apple subcategories could be: +20% saturation in the red channel and increased brightness in the highlight areas.

[0087] For bananas, the product characteristics related to improving appetite can be reflected in: clean peel and no black spots or blemishes. Therefore, improving freshness can enhance the appetite of bananas. The corresponding configuration information for banana subcategories could be: intelligent black spot repair.

[0088] In practical applications, if the specific category of the target product is not identified, or if the specific category of the target product is identified but the corresponding configuration information is not matched, this application embodiment can also provide general configuration information while ensuring basic image quality. For example, general configuration information may include: slight noise reduction, color correction, and fine-tuning of overall contrast to avoid over-resolution distortion and achieve the purpose of enhancing appetite.

[0089] The above is merely an illustrative example, and the embodiments of this application do not specifically limit the configuration information associated with various subcategories.

[0090] In this way, after the first super-resolution model performs super-resolution processing on the text region of the target image and outputs the super-resolution image, the server can also perform image enhancement processing on the image region of the target image according to the second image enhancement method and the second intensity parameter, thereby improving the visual effect of the target image in terms of appetite.

[0091] In one approach, the server can perform image enhancement processing on the text region and the image region of the target image in two separate enhancement processes, sequentially. For example, it can... Figure 3 Based on the scheme shown, the text region is enhanced according to the first image enhancement method and the first intensity parameter to obtain the target image after image enhancement. Then, the image region in the target image after image enhancement is enhanced according to the second image enhancement method and the second intensity parameter. The image obtained after the two enhancement processes is used as the target image after image enhancement returned to the client.

[0092] In another approach, after obtaining the image processing solutions for the two regions, the server can perform differentiated enhancement processing on the text region and the image region in one enhancement process, and return the enhanced target image to the client.

[0093] Preferred Option 2 Based on the image quality assessment results, secondary super-resolution processing can be performed to optimize the display effect of image areas by increasing the number of pixels.

[0094] Specifically, the target image can be first evaluated for image quality, and then a corresponding image processing scheme can be generated based on the evaluation results through intelligent decision-making.

[0095] This application provides a zonal evaluation scheme based on multi-dimensional comprehensive quality assessment. Specifically, it obtains first evaluation dimension information and corresponding first weight information associated with text regions, and second evaluation dimension information and corresponding second weight information associated with subdivided categories; it performs image quality assessment on text regions based on the first evaluation dimension information and first weight information to obtain a first evaluation score, and performs image quality assessment on image regions based on the second evaluation dimension information and second weight information to obtain a second evaluation score; and it obtains the evaluation result of the target image based on the first evaluation score and the second evaluation score.

[0096] As an example, the evaluation dimension information may include sharpness, noise, contrast, edge sharpness, etc. The evaluation dimensions associated with the text area and the sub-category may be the same or different, depending on the usage requirements. This application embodiment does not limit this.

[0097] Taking sharpness and noise as the first and second evaluation dimensions, respectively, as examples, when evaluating the quality of the pork belly image mentioned above, for the text region, readability is emphasized more, so a higher weight can be configured for sharpness and a relatively lower weight for noise. This means sharpness has a greater impact on the quality evaluation results for the text region. For the pork belly image region, visual effect is emphasized more, and excessive noise should be avoided to prevent it from obscuring muscle texture. Therefore, a higher weight can be configured for noise and a relatively lower weight for sharpness. This means noise has a greater impact on the quality evaluation results for the image region within the pork belly sub-category. The evaluation dimensions used for quality evaluation and the corresponding weight values ​​for each dimension can be determined according to usage requirements. This embodiment is only an example.

[0098] In this example, after obtaining the first and second evaluation scores, their average can be used as the overall evaluation score of the target image to obtain the evaluation result. Alternatively, the weights of the two evaluation scores can be configured according to usage requirements, and their weighted sum can be used as the overall evaluation score of the target image. As an example, the weights of the two evaluation scores can be determined based on the proportion of the text area and the image area in the image. For instance, if the text area occupies a larger proportion, it can be determined that the target image focuses more on evaluating and displaying text information, so a higher weight can be assigned to the first evaluation score; conversely, a higher weight can be assigned to the second evaluation score.

[0099] After obtaining the quality assessment results of the target image, the embodiments of this application can provide the following two specific solutions based on the assessment results, which are illustrated below.

[0100] 1. Schemes requiring over-resolution processing If the evaluation results indicate that the target image has low image quality, it can be determined that super-resolution processing is required. A third image processing scheme is then generated through intelligent decision-making. This third image processing scheme may include the model identifier and parameters of the second super-resolution model. The server can then use the super-resolution processed image output by the first super-resolution model and the second model parameters as input to call the second super-resolution model for super-resolution reconstruction. The output of the second super-resolution model is then used as the super-resolution processed image for image enhancement. Understandably, image enhancement can be performed on the text region of the super-resolution processed image output by the second super-resolution model, or on other image regions as needed; the specific process is described above.

[0101] The second super-resolution model in this example can be a general super-resolution model; or, if corresponding super-resolution models are trained for different sub-categories, the model identifier of the second super-resolution model for super-resolution processing can be determined by comprehensively considering the sub-categories of the target product. Similarly, a general super-resolution model or a category-specific super-resolution model can be trained based on the capabilities of traditional models or AI models, and this application embodiment does not limit this.

[0102] For the second model parameters, the super-resolution factor can be flexibly configured according to usage requirements. For example, it can be configured by default to perform 2x super-resolution. Alternatively, at least one evaluation score threshold can be set to perform more granular differentiation of the evaluation results. For example, if the overall image evaluation score is below the evaluation score threshold, 4x super-resolution can be performed; if it is not below the evaluation score threshold, 2x super-resolution can be performed. This application embodiment does not specifically limit this. The image block information can be reflected in cutting the target image into a preset number of image blocks, such as cutting it into 4 blocks, so that the second super-resolution model can perform super-resolution reconstruction on each image block separately. The overlapping pixel value can be reflected as 64 pixels.

[0103] Understandably, if the second super-resolution model performs image block segmentation during the super-resolution process, then after performing super-resolution processing and image enhancement processing on the image blocks respectively, image block fusion can be performed, and the fused image can be used as the target image after intelligent image processing.

[0104] 2. Solutions that do not require over-resolution processing: If the evaluation results indicate that the target image has high image quality, then it can be determined that no super-resolution processing is needed for the target image. Meanwhile, to improve the efficiency of the model's super-resolution processing, this application aims to compress the image region of the target image without image distortion, and uses the compressed target image as input to the first super-resolution model, which then performs super-resolution processing on the text region.

[0105] Preferred Option 3 This application embodiment can also combine the above-mentioned preferred solutions one and two to provide a preferred solution three. Under this solution, the server can evaluate the image quality of the target image, perform secondary super-resolution processing on the target image based on the evaluation results, and identify the sub-categories of the target goods in the target image, and perform image enhancement processing on the image regions based on the sub-categories. Specific processes can be found in the above description and will not be illustrated here.

[0106] As described above, this application can optimize the image processing of target images containing text regions. In practical applications, the target image may not contain text information. Correspondingly, when the server determines that the target product is related to the food category and the target image contains food-related product image content, it can identify the sub-category of the target product under the food category and perform image quality assessment on the target image. Furthermore, based on image processing knowledge information related to enhancing the appetizing expressiveness of the image associated with the sub-category and the evaluation results of the image quality assessment, a fourth image processing scheme for the target image is generated. The fourth image processing scheme includes: a third image enhancement method and corresponding third intensity parameters, and a model identifier and third model parameters of a third super-resolution model determined when the evaluation result indicates that super-resolution processing of the target image is required. In this way, the server can take the target image and the third model parameters as input, call the third super-resolution model, and the third super-resolution model performs super-resolution reconstruction on the target image, specifically the entire target image, according to the third model parameters, and outputs the super-resolution processed image. Then, according to the third image enhancement method and the third intensity parameter, the super-resolution processed image is enhanced to obtain the target image after intelligent image processing, so as to publish product evaluation information based on it.

[0107] Understandably, this example performs optimization processing in two stages: super-resolution and image enhancement, targeting the entire target image. Similarly, if the evaluation results indicate that super-resolution processing of the target image is unnecessary, lossless compression of the entire image can also be performed. Other aspects of this example can be found in the above description and will not be elaborated upon here.

[0108] In summary, the embodiments of this application can determine the target image to be processed from the product reviews submitted by users for the target product, and provide image optimization processing for the target image before the review is published, so as to ensure the high-quality publication of the product review image.

[0109] Specifically, when it is determined that the image content generated from the outer packaging of the target product includes text information on the outer packaging, the server can generate a first image processing scheme for the text area where the text information is located.

[0110] The image processing procedure of this application can be implemented in two stages. First, the text region can be super-resolution processed using a super-resolution model to optimize image quality by increasing the number of pixels. Second, the text region can be further enhanced to improve image quality by increasing pixel quality. Therefore, the first image processing scheme generated by the server can include: a model identifier and first model parameters of the first super-resolution model used for super-resolution processing; and a first image enhancement method and corresponding first intensity parameters used for image enhancement processing.

[0111] The server performs super-resolution model calls and image enhancement processing according to the first image processing scheme to obtain the target image after intelligent image processing, which can be used when publishing product evaluation information.

[0112] This approach helps improve the image quality of target images, thereby enhancing the reference value of user-submitted product reviews. For users submitting product reviews, it helps identify their text and image reviews as high-quality and gives them special markings, improving the user experience of posting product reviews. For users viewing product reviews, high-quality text and image reviews help improve the browsing experience and facilitate quick purchasing decisions, thus increasing user shopping efficiency and the store's product conversion rate.

[0113] As another aspect of this application, a smart image processing scheme aimed at enhancing appetite can also be provided. See [link to relevant documentation]. Figure 4 The flowchart shown may include: S401: Receive product review content submitted by the user for the target product, and determine the target image to be processed from the product review content; the target image includes image content generated by taking a picture of the outer packaging of the target product; S402: If the target product is related to the food category, and the target image includes product image content related to the food, then identify the sub-category of the target product under the food category, and obtain image processing knowledge information related to the sub-category and improving the appetite-enhancing effect of the image. S403: Generate an image optimization scheme based on the knowledge information, and perform intelligent image processing on the target image so as to publish product evaluation information based on the intelligently processed target image.

[0114] In other words, this application can perform targeted image optimization processing with the core objective of enhancing appetite. Specifically, if the target product is related to the food category, the sub-category of the target product under the food category can be identified based on the product image content of the target image. An image optimization scheme can be generated based on the image processing knowledge information related to appetite enhancement associated with this sub-category. Then, targeted intelligent image processing can be performed on the target image according to the image optimization scheme to achieve the purpose of enhancing the visual appetite of the target image.

[0115] The implementation process of the relevant content in this embodiment can be found in the above description, and will not be illustrated here.

[0116] In addition, this application's solution can not only obtain product reviews submitted by users for target products and optimize the target images before review distribution, but also optimize images submitted for publication from other user-generated content (UGC) pages provided by the product information service system application.

[0117] Specifically, the client receives user-generated content (UGC) posted by a user for a target product and sends it to the server. The server then determines the target image to be processed from the UGC, which includes user experience information related to the target product. The server generates a first image processing scheme for the text area containing the user experience information, and uses a first super-resolution model to perform super-resolution reconstruction on the text area according to the first image processing scheme. The server then performs image enhancement processing on the super-resolution processed image to obtain the intelligently processed target image, so that the UGC can be posted based on the intelligently processed target image.

[0118] In other words, when users publish content through the UGC page of an application that provides online shopping services, the application client can send the user-generated content published by the user for the target product to the server. The server then determines the target image to be processed from the content, optimizes the target image according to the solution in this application, and then publishes the content.

[0119] Similarly, image optimization can be implemented without the user's awareness, or image processing can be initiated based on the user's choice. The process of client-server collaboration in image optimization is explained above and will not be illustrated here.

[0120] In this example, user experience information related to the target product can be reflected in the user experience of the target product, such as user experience information related to the user tutorial of the target product published by the user through the UGC page; or, it can be reflected in the evaluation information of the target product, such as user experience information related to the actual test feedback of the target product published by the user through the UGC page; or, it can be reflected in the processing and production information of the target product, such as user experience information related to the creative matching of the target product published by the user through the UGC page.

[0121] If user-generated content is published in the form of images, videos, GIFs, etc., the client can send the user-generated content to the server, which will then determine the target image and optimize the image for the text area containing the experience information included in the target image.

[0122] For example, in a creative recipe presentation using a target product as the ingredient, the text area containing the experiential information related to the food preparation method within the target image can be optimized according to the solution in this application. Furthermore, the target image may also include product images related to the ingredients used in the dish preparation; similarly, the image area containing these product images can be optimized according to the solution in this application to enhance the appetizing appeal of the target image after intelligent image processing. The specific implementation process can be found above and will not be detailed here.

[0123] It should be noted that the embodiments of this application may involve the use of user data. In practical applications, user-specific personal data may be used in the scheme described herein within the scope permitted by applicable laws and regulations, provided that it complies with the applicable laws and regulations of the country (e.g., with the user's explicit consent, with the user being properly notified, etc.).

[0124] Corresponding to the foregoing method embodiments, this application also provides an intelligent image processing device applied to a server. See also Figure 5 The device may include: The target image determination unit 501 is used to receive product evaluation content submitted by a user for a target product, and determine the target image to be processed from the product evaluation content; the target image includes image content generated by taking a picture of the outer packaging of the target product, and the image content includes the text area where the text information on the outer packaging is located; The image processing scheme generation unit 502 is used to generate a first image processing scheme for the text region. The first image processing scheme includes: a model identifier and first model parameters of a first super-resolution model, and a first image enhancement method and a corresponding first intensity parameter. The super-resolution model calling unit 503 is used to call the first super-resolution model so that the first super-resolution model can perform super-resolution reconstruction of the text region of the target image according to the first model parameters and output the super-resolution processed image. The image enhancement processing unit 504 is used to perform image enhancement processing on the super-resolution processed image according to the first image enhancement method and the first intensity parameter, so as to publish product evaluation information based on the target image after intelligent image processing.

[0125] Wherein, if the target product is related to the food category, the target image also includes product image content related to the food, and the device further includes: A sub-category identification unit is used to identify the sub-category of the target product under the food category; The image processing scheme generation unit can also be used to: obtain image processing knowledge information related to enhancing the appetite expression of the image associated with the subdivided product category, and generate a second image processing scheme for the image area where the product image is located based on the knowledge information. The second image processing scheme includes a second image enhancement method and a corresponding second intensity parameter. The image enhancement processing unit can also be used to: after obtaining the super-resolution processed image, perform image enhancement processing on the image region according to the second image enhancement method and the second intensity parameter.

[0126] The device further includes: An image quality assessment unit is used to assess the image quality of the target image. The image processing scheme generation unit can also be used to: generate a third image processing scheme for the target image when the evaluation result indicates that the target image needs to be super-resolution processed, wherein the third image processing scheme includes the model identifier of the second super-resolution model and the second model parameters; The super-resolution model invocation unit can also be used to: invoke the second super-resolution model so that the second super-resolution model performs super-resolution reconstruction on the super-resolution processed image output by the first super-resolution model according to the second model parameters, and uses the output of the second super-resolution model as the super-resolution processed image for image enhancement processing.

[0127] The device further includes: The compression processing unit is configured to compress the image region with the goal of preserving image quality when the evaluation result indicates that super-resolution processing of the target image is not required, and to use the compressed target image as input to the first super-resolution model for super-resolution processing.

[0128] Specifically, the image quality assessment unit can be used for: Obtain the first evaluation dimension information and the corresponding first weight information associated with the text region, and the second evaluation dimension information and the corresponding second weight information associated with the subdivided category; The text region is evaluated for image quality based on the first evaluation dimension information and the first weight information to obtain a first evaluation score. The image region is evaluated for image quality based on the second evaluation dimension information and the second weight information to obtain a second evaluation score. The evaluation result of the target image is obtained based on the first evaluation score and the second evaluation score.

[0129] The device further includes: The sharpening processing unit is used to sharpen the text region of the target image and use the sharpened target image as input to the first super-resolution model for super-resolution processing.

[0130] The first model parameter includes image block information, which is used to indicate that the target image is not cut into image blocks.

[0131] The second model parameters include image block information, which indicates that the target image is cut into a preset number of image blocks so that the second super-resolution model can perform super-resolution reconstruction on the image blocks respectively. The device further includes: The image block fusion unit is used to perform super-resolution processing and image enhancement processing on the image blocks respectively, and then perform image block fusion to obtain the target image after intelligent image processing.

[0132] If the target image does not include the text region, the device further includes: The relevant information acquisition unit is used to determine that the target product is related to the food category, and that the target image includes product image content related to the food, identify the sub-category of the target product under the food category, and perform image quality assessment on the target image; The image processing scheme generation unit can also be used to: generate a fourth image processing scheme for the target image based on the image processing knowledge information related to enhancing the appetite expression of the image associated with the subdivided category and the evaluation result of the image quality assessment. The fourth image processing scheme includes: a third image enhancement method and a corresponding third intensity parameter, and a model identifier and third model parameters of a third super-resolution model determined when the evaluation result indicates that the target image needs to be super-resolution processed. The super-resolution model invocation unit can also be used to: invoke the third super-resolution model so that the third super-resolution model can perform super-resolution reconstruction of the target image according to the parameters of the third model and output the super-resolution processed image. The image enhancement processing unit can also be used to: perform image enhancement processing on the super-resolution processed image according to the third image enhancement method and the third intensity parameter, so as to publish product evaluation information based on the target image after intelligent image processing.

[0133] The device further includes: An adjustment request receiving unit is used to receive the image adjustment request submitted by the user for the target image; An image adjustment unit is used to determine the image adjustment method and adjustment parameters corresponding to the image adjustment requirements, and to perform image adjustment on the target image so as to publish product evaluation information based on the image-adjusted target image.

[0134] Corresponding to the foregoing method embodiments, this application also provides an intelligent image processing device applied to a server. See also Figure 6 The device may include: The target image determination unit 601 is used to receive product evaluation content submitted by a user for a target product, and determine the target image to be processed from the product evaluation content; the target image includes image content generated by taking a picture of the outer packaging of the target product; The sub-category identification unit 602 is used to identify the sub-category of the target product under the food category, and to obtain image processing knowledge information related to the sub-category and related to improving the appetite expression of the image, when the target product is related to the food category and the target image includes product image content related to the food. The image optimization scheme generation unit 603 is used to generate an image optimization scheme based on the knowledge information and perform intelligent image processing on the target image so as to publish product evaluation information based on the intelligently processed target image.

[0135] Corresponding to the foregoing method embodiments, this application also provides an intelligent image processing device applied to a client. The device may include: A product review content sending unit is used to receive product review content submitted by users for a target product and send it to the server so that the server can determine the target image to be processed from the product review content. The target image includes image content generated by taking a picture of the outer packaging of the target product, and the image content includes a text area containing text information on the outer packaging. A first image processing scheme is generated for the text area, and a first super-resolution model is called according to the first image processing scheme to perform super-resolution reconstruction of the text area, and image enhancement processing is performed on the super-resolution processed image so as to publish product review information based on the target image after intelligent image processing.

[0136] The device further includes: The first operation option providing unit is used to receive product evaluation content submitted by the user for the target product and provide a first operation option for starting intelligent image processing; The product review content sending unit can be specifically used to: after obtaining an image processing request for the target image in the product review content through the first operation option, send the product review content to the server.

[0137] The device further includes: The second operation option providing unit is used to provide a second operation option for submitting image adjustment requests; The image adjustment request acquisition unit is used to obtain the image adjustment request submitted by the user for the target image through the second operation option, and send it to the server so that the server can determine the image adjustment method and adjustment parameters corresponding to the image adjustment request, perform image adjustment on the target image, and obtain the adjusted target image.

[0138] The second operation option consists of tabs provided for different image adjustment methods. The image adjustment request acquisition unit can be specifically used to: when the target tab is selected, determine the image adjustment method represented by the target tab as the image adjustment request submitted by the user.

[0139] The second operation option is an information input box. The image adjustment request acquisition unit can be specifically used to: obtain the image adjustment request submitted by the user through the information input box.

[0140] Corresponding to the foregoing method embodiments, this application also provides an intelligent image processing device applied to an application client of a commodity information service system. The device may include: The user-generated content sending unit is used to receive user-generated content published by a user for a target product and send it to the server so that the server can determine the target image to be processed from the user-generated content. The target image includes user experience information related to the target product. The unit generates a first image processing scheme for the text area where the user experience information is located, and calls a first super-resolution model to perform super-resolution reconstruction on the text area according to the first image processing scheme. The unit also performs image enhancement processing on the super-resolution processed image to obtain a target image after intelligent image processing, so that user-generated content can be published based on the target image after intelligent image processing.

[0141] In addition, embodiments of this application also provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method described in any of the foregoing method embodiments.

[0142] And an electronic device, comprising: One or more processors; and A memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform the steps of the method described in any of the foregoing method embodiments.

[0143] A computer program product includes a computer program / computer executable instructions that, when executed by a processor in an electronic device, implement the steps of the method described in the foregoing method embodiments.

[0144] in, Figure 7 The architecture of an electronic device is illustrated by example. For instance, device 700 could be a mobile phone, computer, digital broadcasting terminal, messaging device, game console, tablet device, medical device, fitness equipment, personal digital assistant, aircraft, etc.

[0145] Reference Figure 7 The device 700 may include one or more of the following components: a processing component 702, a memory 704, a power supply component 706, a multimedia component 708, an audio component 710, an input / output (I / O) interface 712, a sensor component 714, and a communication component 716.

[0146] Processing component 702 typically controls the overall operation of device 700, such as operations associated with display, telephone calls, data communication, camera operation, and recording. Processing component 702 may include one or more processors 720 to execute instructions to perform all or part of the steps of the methods provided in this disclosure. Furthermore, processing component 702 may include one or more modules to facilitate interaction between processing component 702 and other components. For example, processing component 702 may include a multimedia module to facilitate interaction between multimedia component 708 and processing component 702.

[0147] Memory 704 is configured to store various types of data to support the operation of device 700. Examples of this data include instructions for any application or method operating on device 700, contact data, phonebook data, messages, pictures, videos, etc. Memory 704 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0148] Power supply component 706 provides power to various components of device 700. Power supply component 706 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to device 700.

[0149] Multimedia component 708 includes a screen that provides an output interface between device 700 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may sense not only the boundaries of touch or swipe actions but also the duration and pressure associated with the touch or swipe operation. In some embodiments, multimedia component 708 includes a front-facing camera and / or a rear-facing camera. When device 700 is in an operating mode, such as a shooting mode or a video mode, the front-facing camera and / or rear-facing camera may receive external multimedia data. Each front-facing camera and rear-facing camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

[0150] Audio component 710 is configured to output and / or input audio signals. For example, audio component 710 includes a microphone (MIC) configured to receive external audio signals when device 700 is in an operating mode, such as call mode, recording mode, and voice recognition mode. The received audio signals may be further stored in memory 704 or transmitted via communication component 716. In some embodiments, audio component 710 also includes a speaker for outputting audio signals.

[0151] I / O interface 712 provides an interface between processing component 702 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to, home buttons, volume buttons, power buttons, and lock buttons.

[0152] Sensor assembly 714 includes one or more sensors for providing state assessments of various aspects of device 700. For example, sensor assembly 714 may detect the on / off state of device 700, the relative positioning of components such as the display and keypad of device 700, changes in the position of device 700 or a component of device 700, the presence or absence of user contact with device 700, the orientation or acceleration / deceleration of device 700, and temperature changes of device 700. Sensor assembly 714 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 714 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, sensor assembly 714 may also include an accelerometer, a gyroscope, a magnetometer, a pressure sensor, or a temperature sensor.

[0153] Communication component 716 is configured to facilitate wired or wireless communication between device 700 and other devices. Device 700 can access wireless networks based on communication standards, such as WiFi, or mobile communication networks such as 2G, 7G, 4G / LTE, and 7G. In one exemplary embodiment, communication component 716 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, communication component 716 also includes a near-field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0154] In an exemplary embodiment, device 700 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the methods described above.

[0155] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 704 including instructions, which can be executed by a processor 720 of device 700 to perform the method provided by the present disclosure. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0156] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0157] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for system or system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and relevant parts can be referred to the descriptions in the method embodiments. The systems and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0158] The solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. An intelligent image processing method, characterized in that, Applied to the server side, the method includes: The system receives product reviews submitted by users for a target product and determines a target image to be processed from the product review content. The target image includes image content generated by taking a picture of the outer packaging of the target product, and the image content includes the text area where the text information on the outer packaging is located. A first image processing scheme is generated for the text region. The first image processing scheme includes: a model identifier and first model parameters of a first super-resolution model, and a first image enhancement method and a corresponding first intensity parameter. The first super-resolution model is invoked so that the first super-resolution model performs super-resolution reconstruction of the text region of the target image according to the first model parameters and outputs the super-resolution processed image. Based on the first image enhancement method and the first intensity parameter, the super-resolution processed image is subjected to image enhancement processing so as to publish product evaluation information based on the target image after intelligent image processing.

2. The method according to claim 1, characterized in that, If the target product is related to the food category, the target image also includes product image content related to the food, and the method further includes: Identify the sub-category of the target product under the food category, obtain image processing knowledge information related to the sub-category and related to enhancing the appetite expression of the image, and generate a second image processing scheme for the image region where the product image is located based on the knowledge information. The second image processing scheme includes a second image enhancement method and a corresponding second intensity parameter. After obtaining the super-resolution processed image, image enhancement processing is performed on the image region according to the second image enhancement method and the second intensity parameter.

3. The method according to claim 2, characterized in that, The method further includes: The target image is subjected to image quality assessment, and if the assessment result indicates that the target image needs to be super-resolution processed, a third image processing scheme for the target image is generated. The third image processing scheme includes the model identifier of the second super-resolution model and the second model parameters. The second super-resolution model is invoked so that the second super-resolution model performs super-resolution reconstruction on the super-resolution processed image output by the first super-resolution model according to the second model parameters, and the output of the second super-resolution model is used as the super-resolution processed image for image enhancement processing.

4. The method according to claim 3, characterized in that, The method further includes: If the evaluation result indicates that super-resolution processing of the target image is not required, the image region is compressed with the goal of maintaining image integrity, and the compressed target image is used as the input of the first super-resolution model for super-resolution processing.

5. The method according to claim 3 or 4, characterized in that, The image quality assessment of the target image includes: Obtain the first evaluation dimension information and the corresponding first weight information associated with the text region, and the second evaluation dimension information and the corresponding second weight information associated with the subdivided category; The text region is evaluated for image quality based on the first evaluation dimension information and the first weight information to obtain a first evaluation score. The image region is evaluated for image quality based on the second evaluation dimension information and the second weight information to obtain a second evaluation score. The evaluation result of the target image is obtained based on the first evaluation score and the second evaluation score.

6. The method according to any one of claims 1 to 4, characterized in that, The method further includes: The text region of the target image is sharpened, and the sharpened target image is used as the input of the first super-resolution model for super-resolution processing.

7. The method according to any one of claims 1 to 4, characterized in that, If the target image does not include the text region, the method further includes: The system determines that the target product is related to the food category, and that the target image includes product image content related to the food. It also identifies the sub-category of the target product under the food category and performs image quality assessment on the target image. Based on the image processing knowledge information related to enhancing the appetite expression of the image associated with the subdivided category and the evaluation results of the image quality assessment, a fourth image processing scheme is generated for the target image. The fourth image processing scheme includes: a third image enhancement method and a corresponding third intensity parameter, and a model identifier and third model parameters of a third super-resolution model determined when the evaluation result indicates that the target image needs to be super-resolution processed. The third super-resolution model is invoked so that the third super-resolution model can perform super-resolution reconstruction of the target image according to the parameters of the third model and output the super-resolution processed image. Based on the third image enhancement method and the third intensity parameter, the super-resolution processed image is subjected to image enhancement processing so as to publish product evaluation information based on the target image after intelligent image processing.

8. The method according to any one of claims 1 to 4, characterized in that, The method further includes: If a user submits an image adjustment request for the target image, the system determines the image adjustment method and parameters corresponding to the request, and adjusts the target image accordingly, so as to publish product review information based on the adjusted target image.

9. An intelligent image processing method, characterized in that, Applied to the server side, the method includes: The system receives product reviews submitted by users for a target product and determines the target image to be processed from the product reviews; the target image includes an image generated by taking a picture of the outer packaging of the target product. If the target product is related to the food category, and the target image includes product image content related to the food, then the sub-category of the target product under the food category is identified, and image processing knowledge information related to the sub-category and improving the appetite-enhancing effect of the image is obtained. An image optimization scheme is generated based on the knowledge information, and intelligent image processing is performed on the target image so as to publish product evaluation information based on the intelligently processed target image.

10. An intelligent image processing method, characterized in that, Applied to a client, the method includes: The system receives product reviews submitted by users for a target product and sends them to the server. The server then determines the target image to be processed from the product review content. The target image includes an image of the outer packaging of the target product, and the image content includes a text area containing text information on the outer packaging. A first image processing scheme is generated for the text area, and a first super-resolution model is called to perform super-resolution reconstruction of the text area according to the first image processing scheme. Image enhancement processing is then performed on the super-resolution processed image so that product review information can be published based on the intelligently image-processed target image.

11. The method according to claim 10, characterized in that, The method further includes: Receive product reviews submitted by users for the target product and provide a first operation option for initiating intelligent image processing; After obtaining an image processing request for the target image in the product review content through the first operation option, the product review content is sent to the server.

12. The method according to claim 10, characterized in that, The method further includes: Provide a second operation option for submitting image adjustment requests; The second operation option obtains the image adjustment request submitted by the user for the target image and sends it to the server so that the server can determine the image adjustment method and adjustment parameters corresponding to the image adjustment request, perform image adjustment on the target image, and obtain the adjusted target image.

13. An intelligent image processing method, characterized in that, The method, which is applied to an application client for a commodity information service system, includes: The system receives user-generated content (UGC) posted by a user for a target product and sends it to the server. The server then determines the target image to be processed from the UGC, where the target image includes user experience information related to the target product. A first image processing scheme is generated for the text region containing the user experience information. Based on the first image processing scheme, a first super-resolution model is invoked to perform super-resolution reconstruction on the text region. Image enhancement processing is then performed on the super-resolution image to obtain a target image after intelligent image processing. User-generated content is then posted based on this intelligent image-processed target image.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program performs the steps of the method described in any one of claims 1 to 13.

15. An electronic device, characterized in that, include: One or more processors; as well as A memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform the steps of the method according to any one of claims 1 to 13.

16. A computer program product comprising a computer program / computer-executable instructions, characterized in that, When the computer program / computer-executable instructions are executed by a processor in an electronic device, they implement the steps of the method according to any one of claims 1 to 13.

Images