Methods, devices, storage media, and program products for displaying model images
By pre-labeling a set of images containing models and merchant logos, and selecting the target image for display, the problem of duplicate display of images of the same model in online home design platforms is solved, improving user selection efficiency and experience.
Patent Information
- Application Number
- CN202310004914.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-01-03
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2043-01-03
AI Technical Summary
On online home design platforms, images of the same model are displayed repeatedly on the same page, leading to low user selection efficiency and a reduced user experience.
By pre-labeling model images with model and merchant identifiers, and selecting a target image from a set of similar images to display when a user searches, the duplicate display of the same image on the same page is avoided. The target image is determined by using hash features and frequency filtering algorithms.
It improves the efficiency of model image display, avoids the repeated display of the same model image, and enhances the user browsing experience.
Smart Images

Figure CN115935093B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of data processing technology, and in particular to a method, device, storage medium and program product for displaying model images. Background Technology
[0002] With the rapid development of internet technology, online home design platforms have emerged, providing users with a selection of home design options. These options are typically displayed as 3D model images on the user's screen, allowing for a direct view of the design's effect. For example, multiple home model images can be displayed on a single browsing page, enabling users to choose according to their specific needs.
[0003] With the development of online home design platforms, in order to improve the user experience, merchants can upload images of their own home furnishing models. When other users search for related home furnishings, these models can be displayed on the page. However, in reality, when displaying model images, a single page often shows many similar or identical home furnishing models. Displaying these similar or identical models together on the same page reduces the efficiency of users selecting home furnishing models and creates a poor user experience. Summary of the Invention
[0004] The main objective of this application is to provide a method, device, storage medium, and program product for displaying model images, which ensures that images of the same model are not displayed repeatedly on the same page, avoids the inefficiency caused by displaying the same object model image multiple times on the same page, improves the display effect of object model images, and enhances the user experience.
[0005] In a first aspect, embodiments of this application provide a method for displaying model images, comprising: responding to a user's search command, acquiring a plurality of images matching the search command, each image including an object model matching the search command; determining whether a set of images of the same model exists among the plurality of images, wherein the object models of each image in the set of images of the same model are marked as the same model; if the set of images of the same model exists among the plurality of images, determining a target image from the set of images of the same model; and displaying the target image and non-identical images on a page, wherein the non-identical images are images among the plurality of images excluding the set of images of the same model.
[0006] In one embodiment, determining whether a set of images with the same model exists among the plurality of images includes: determining whether a set of images with the same model identifier exists among the plurality of images; if a set of images with the same model identifier exists among the plurality of images, determining that the set of images with the same model exists among the plurality of images; otherwise, determining that the set of images with the same model does not exist among the plurality of images.
[0007] In one embodiment, determining a target image from the set of images with the same model if such a set exists among the plurality of images includes: if such a set of images with the same model exists among the plurality of images, determining whether there is a set of images with the same merchant identifier in the set of images with the same model; if such a set of images with the same merchant exists in the set of images with the same model, determining a candidate image with the same model from the set of images with the same merchant; comparing the usage frequency of a first remaining image with the same model in the set of images with the usage frequency of the candidate image with the usage frequency of the image with the same model, and determining the image with the highest usage frequency as the target image, wherein the first remaining image with the same model is an image in the set of images with the same model excluding the set of images with the same merchant.
[0008] In one embodiment, if the set of images of the same model exists, determining a candidate image of the same model from the set of images of the same merchant includes: if the set of images of the same model exists, determining the image with the highest frequency of use in the set of images of the same merchant as the candidate image of the same model.
[0009] In one embodiment, the same merchant image set is a set of images of the same object model in different colors; after displaying the target image and the non-same image on the page, the method further includes: if the same merchant image set exists in the same model image set, and the candidate same image is determined as the target image, the second remaining same image in the same merchant image set is configured to be displayed on the next level page of the current page, and the second remaining same image is the image in the same merchant image set excluding the candidate same image.
[0010] In one embodiment, if the set of images of the same model exists among the plurality of images, determining a target image from the set of images of the same model further includes: if the set of images of the same merchant does not exist in the set of images of the same model, determining the image of the same model that is used most frequently in the set of images of the same model as the target image.
[0011] In one embodiment, if the set of images of the same model exists among the plurality of images, after displaying the target image and the non-same model images on the page, the method further includes: if the set of images of the same merchant does not exist in the set of images of the same model, configuring the third remaining images of the same model in the set of images of the same model on different pages for display, wherein the third remaining images of the same model are on different pages from the images in the set of images of the same model excluding the target image.
[0012] In one embodiment, before obtaining multiple images matching the user's search command, the method further includes: obtaining multiple original images of an object model to be processed; extracting image features from each of the original images; and marking two original images whose image features have a similarity greater than a preset threshold as the same type.
[0013] In one embodiment, the step of extracting image features from each of the original images includes: calculating hash features for each of the original images, wherein the hash features include one or more of an average hash value, a perceptual hash value, and a differential hash value; and generating image features for each of the original images based on the hash features.
[0014] In one embodiment, the step of extracting image features from each of the original images further includes: calculating the color mean of each channel in each of the original images, and generating image features for each of the original images based on the color mean and the hash feature.
[0015] Secondly, embodiments of this application provide a method for displaying model images, comprising: responding to a user's search command for a home furnishing model, acquiring a plurality of images matching the search command, each image including a home furnishing model matching the search command; determining whether a set of images of the same model exists among the plurality of images, wherein the home furnishing models in each image of the set of images of the same model are marked as the same model; if the set of images of the same model exists among the plurality of images, determining a target image from the set of images of the same model; and displaying the target image and non-identical images on a page, wherein the non-identical images are images among the plurality of images excluding the set of images of the same model.
[0016] Thirdly, embodiments of this application provide a display device for a model image, comprising:
[0017] The first acquisition module is used to acquire multiple images that match the user's search command in response to the user's search command, and each image includes an object model that matches the search command.
[0018] The judgment module is used to determine whether there is a set of images with the same model among the multiple images, and the object models of each image in the set of images with the same model are marked as the same.
[0019] The determining module is used to determine a target image from the set of images of the same model if the set of images of the same model exists among the plurality of images;
[0020] The display module is used to display the target image and non-identical images on the page. The non-identical images are images other than the set of identical model images among the multiple images.
[0021] In one embodiment, the determination module is used to determine whether there is an image set with the same model identifier among the plurality of images; if there is an image set with the same model identifier among the plurality of images, it is determined that the same model image set exists among the plurality of images; otherwise, it is determined that the same model image set does not exist among the plurality of images.
[0022] In one embodiment, the determining module is configured to: if the set of images with the same model exists among the plurality of images, determine whether there is a set of images with the same merchant identifier in the set of images with the same model; if the set of images with the same merchant exists in the set of images with the same model, determine a candidate image with the same model from the set of images with the same merchant; compare the usage frequency of the first remaining image with the same model in the set of images with the usage frequency of the candidate image with the usage frequency of the image with the same model, and determine the image with the highest usage frequency as the target image, wherein the first remaining image with the same model is an image in the set of images with the same model excluding the set of images with the same merchant.
[0023] In one embodiment, the determining module is used to determine the image with the highest frequency of use in the same merchant image set as the candidate same product image if the same merchant image set exists in the same model image set.
[0024] In one embodiment, the same merchant image set is a set of images of the same object model in different colors; after displaying the target image and the non-same image on the page, the display module is further configured to, if the same merchant image set exists in the same model image set and the candidate same image is determined as the target image, configure the second remaining same image in the same merchant image set to be displayed on the next level page of the current page, wherein the second remaining same image is an image in the same merchant image set excluding the candidate same image.
[0025] In one embodiment, the determining module is configured to determine the most frequently used image in the same model image set as the target image if the same merchant image set does not exist in the same model image set.
[0026] In one embodiment, if the set of images of the same model exists among the plurality of images, after displaying the target image and the non-same model images on the page, the display module is further configured to, if the set of images of the same model does not exist among the set of images of the same merchant, configure the third remaining images of the same model in the set of images of the same model on different pages for display, wherein the third remaining images of the same model are on different pages from the images in the set of images of the same model excluding the target image.
[0027] In one embodiment, the system further includes: a second acquisition module, configured to acquire multiple original images of an object model to be processed before acquiring multiple images matching the search command in response to the user's search command; an extraction module, configured to extract image features from each of the original images respectively; and a labeling module, configured to label two original images whose image features have a similarity greater than a preset threshold as the same.
[0028] In one embodiment, the extraction module is configured to calculate the hash features of each of the original images, the hash features including one or more of the average hash value, perceptual hash value and differential hash value; and generate image features of each of the original images based on the hash features.
[0029] In one embodiment, the extraction module is further configured to calculate the color mean of each channel in each of the original images, and generate image features of each of the original images based on the color mean and the hash feature.
[0030] Fourthly, embodiments of this application provide an electronic device, including:
[0031] At least one processor; and
[0032] A memory that is communicatively connected to the at least one processor;
[0033] The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, cause the electronic device to perform the method described in any of the above aspects.
[0034] Fifthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the method described in any of the above aspects.
[0035] Sixthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the method described in any of the above aspects.
[0036] In a seventh aspect, embodiments of this application provide a method for recognizing images of the same model, comprising: acquiring multiple images to be processed, each image to be processed including an object model with the same attributes; determining the model identifier corresponding to the object model in each image to be processed; determining whether there is an image set with the same model identifier among the multiple images to be processed; if there is an image set with the same model identifier among the multiple images to be processed, determining that the image set is an image of the same model.
[0037] The model image display method, device, storage medium, and program product provided in this application embodiment, when a user searches for a certain object model, obtains multiple images of the object model, and determines whether there is an image set marked as the same model. If so, a target image is determined from the image set of the same object model, and the target image and the image without the same model are displayed on the page. This can ensure that the image of the same model is not displayed repeatedly on the same page, avoid the inefficiency caused by displaying the same object model multiple times, improve the display effect of the object model image, and improve the user browsing experience. Attached Figure Description
[0038] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are some embodiments of the invention, and that those skilled in the art can obtain other drawings based on these drawings without any inventive effort.
[0039] Figure 1 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application;
[0040] Figure 2 A scene diagram illustrating a model image display system provided in an embodiment of this application;
[0041] Figure 3 A flowchart illustrating a method for displaying a model image provided in an embodiment of this application;
[0042] Figure 4 This is a schematic diagram illustrating the display of a model image on a page, as provided in an embodiment of this application.
[0043] Figure 5A A flowchart illustrating a method for displaying a model image provided in an embodiment of this application;
[0044] Figure 5BThis is a schematic diagram illustrating a process for calculating image features from an original image, as provided in an embodiment of this application.
[0045] Figure 6 A flowchart illustrating a method for recognizing images of the same model provided in an embodiment of this application;
[0046] Figure 7 A flowchart illustrating a method for displaying a model image provided in an embodiment of this application;
[0047] Figure 8 This is a schematic diagram of the structure of a model image display device provided in an embodiment of this application.
[0048] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation
[0049] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.
[0050] In this article, the term "and / or" is used to describe the relationship between related objects. Specifically, it means that there can be three kinds of relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, or B exists alone.
[0051] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation portals are provided for users to choose to authorize or refuse.
[0052] To clearly describe the technical solutions of the embodiments of this application, the terms involved in this application are first defined as follows:
[0053] Time complexity: The time complexity of an algorithm is a function that qualitatively describes the algorithm's running time. Time complexity is often expressed using Big O notation, excluding lower-order terms and the leading coefficient of the function. When using this method, the time complexity can be considered asymptotic, that is, it examines the case when the input value approaches infinity.
[0054] aHash: Average hash algorithm.
[0055] pHash: Perceptual hash algorithm.
[0056] dHash: Differential hash algorithm.
[0057] SKU: Stock Keeping Unit, the smallest unit of inventory.
[0058] DCT: Discrete Cosine Transform.
[0059] Fassi: Facebook AI Similarity Search, a similarity search computation library.
[0060] like Figure 1 As shown, this embodiment provides an electronic device 1, including: at least one processor 11 and a memory 12. Figure 1 Taking a processor as an example, the processor 11 and the memory 12 are connected via a bus 10. The memory 12 stores instructions that can be executed by the processor 11. The instructions are executed by the processor 11 to enable the electronic device 1 to perform all or part of the process of the method in the following embodiments, so as to display the same object model image only once on a page, avoid the inefficiency caused by displaying different images of the same object model multiple times on the same page, improve the object model image display effect, and improve the user browsing experience.
[0061] In one embodiment, the electronic device 1 may be a mobile phone, tablet computer, laptop computer, desktop computer, or a large computing system composed of multiple computers.
[0062] Figure 2 This is a schematic diagram of a model image display system 200 provided in an embodiment of this application. For example... Figure 2 As shown, the system includes: a server 210 and a terminal 220, wherein:
[0063] Server 210 can be a data platform that provides image display services for object models, such as an online home design platform. In a real-world scenario, an online home design platform may have multiple servers 210. Figure 2 Taking a single server (210) as an example.
[0064] Terminal 220 can be a computer, mobile phone, tablet, or other device used by the user to log in to the online home design platform. There can also be multiple terminals 220. Figure 2 The following example uses two terminals, 220, for illustration.
[0065] Terminal 220 and server 210 can transmit information via the Internet, enabling terminal 220 to access data on server 210. Both terminal 220 and / or server 210 can be implemented by electronic device 1.
[0066] The model image display method of this application embodiment can be applied to any field that requires the display of model images.
[0067] In online home design scenarios, to enhance the user experience and platform interaction, merchants can upload images of their self-made home furnishing models. When other users search for related home furnishings, these model images can be displayed on the page. For example, if a user searches for a sofa, sofa model images can be displayed for browsing, assisting the user in selecting home furnishings.
[0068] In real-world scenarios, when displaying model images, a single page often shows many similar or identical furniture models. If these similar or identical models are displayed together on the same page, it reduces the efficiency of users selecting furniture models and creates a poor user experience. For example, if a user searches for "three-seater sofa," many images of three-seater sofas will be displayed, with some identical sofa images being shown repeatedly, significantly reducing the efficiency of users selecting a suitable sofa image.
[0069] To address the aforementioned issues, a common approach is to perform a lookup check on the initial searched images to reduce the concentration of similar images. Currently, a common method for lookup detection typically utilizes deep learning techniques to obtain feature vectors from model images. Each time a user searches, the similarity between pairs of model images is calculated; if the similarity falls within a certain range, the images are considered similar. The drawback of this method is that calculating the vector similarity between pairs of model images for each search is computationally complex. This high computational cost is the biggest pain point of pairwise model image comparison methods, resulting in a time complexity of O(n^2). Another approach is local pairwise model image calculation, where model image vectors are indexed using a vector search engine (such as Fassi), and then searched through the index. While this reduces the algorithm's time complexity to O(kn), where k is a constant, the computational cost remains too high when comparing a large number of images. Furthermore, the aforementioned methods generally perform global vector searches, failing to calculate lookup models specific to a particular merchant's model library based on business needs, thus not meeting customized lookup detection requirements.
[0070] To address the aforementioned issues, this application provides a model image display scheme. By pre-marking model images as identical, when a user searches for a specific object model, multiple images of that object model are retrieved, and it is determined whether there is a set of images marked as identical. If so, a target image is selected from the set of identical object model images, and the target image and the image without an identical counterpart are displayed on the page. This ensures that images of identical models are displayed only once on the page, avoiding the inefficiency caused by displaying identical object models multiple times, improving the display effect of object model images, and enhancing the user browsing experience. Furthermore, the computational complexity of the algorithm in this application embodiment can be O(1), meaning the algorithm has constant time, requiring no additional computation and only needing to execute a normal query command, thus reducing algorithm complexity.
[0071] The display scheme for the aforementioned model image can be deployed on server 210, on terminal 220, or partially on server 210 and partially on terminal 220. The choice can be made based on actual needs in a real-world scenario, and this embodiment does not impose any limitations.
[0072] When the display scheme of the model image is deployed entirely or partially on the server 210, the call interface can be opened to the terminal 220 to provide algorithm support to the terminal 220.
[0073] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Where there is no conflict between the embodiments, the following embodiments and features can be combined with each other. Furthermore, the timing of the steps in the following method embodiments is merely an example and not a strict limitation.
[0074] Please refer to Figure 3 This is a method for displaying a model image according to an embodiment of this application. The method can be performed by... Figure 1 The electronic device 1 shown is used to perform this action and can be applied to... Figure 2 In the application scenarios of the model image display system shown, this method aims to avoid the inefficiency of displaying the same object model image multiple times on the same page, improve the display effect of the object model image, and enhance the user browsing experience. This embodiment uses terminal 220 as the execution end as an example, and the method includes the following steps:
[0075] Step 301: In response to the user's search command, acquire multiple images that match the search command, each image including an object model that matches the search command.
[0076] In this step, the search command can be a text command, a voice command, or a gesture command. Taking an online home design platform as an example, when a user designs the interior layout of a room through the platform, they can search for relevant home furnishing materials to assist in selecting home items. Users can log in to the online home design platform via a terminal and then enter the search command, such as through the terminal's interactive interface or via an external keyboard. The search command specifies the image of the object model to be searched. The object model can be a real-world object model, such as a physical sofa in an actual scene, or it can be a virtual object model generated through computer 3D modeling. In the field of home design, the object model can be a model of furniture, home decorations, etc. Correspondingly, multiple images are model images of furniture, home decorations, etc. Home furnishing model images can be provided by the online home design platform's database or uploaded by merchant users. Uploaded model images can be saved in the online home design platform's database. When a user searches for a specific object model, multiple matching images are retrieved from the database. Here, "matching" means that all the searched images contain the object model specified in the search command.
[0077] For example, if a user enters the search command "three-seater sofa" on a home design platform via a terminal, multiple images containing three-seater sofa models will be retrieved from the database.
[0078] Step 302: Determine if a set of images with the same model exists among multiple images. The object models in each image of the set of images with the same model are marked as the same. If yes, proceed to step 303; otherwise, proceed to step 305.
[0079] In this step, in a real-world scenario, the object models in the multiple images matching the search command obtained in step 301 might be the same model. For example, if a total of 100 images of three-seater sofa models are obtained, 30 of them might contain the same three-seater sofa model. If all of these are displayed on the page at once, the same sofa model will appear repeatedly when the user browses a page of images, which is not conducive to the user quickly selecting a suitable sofa model. Therefore, to improve the efficiency of the user browsing model images, the images can be filtered and arranged before displaying them. Specifically, images containing the same model of object model are marked in advance in the database. After obtaining multiple images matching the search command in step 301, it is first determined whether there is a set of images of the same model marked as the same model among the multiple images obtained. If so, proceed to step 303; otherwise, proceed to step 305.
[0080] In one embodiment, step 302 may specifically include: determining whether there exists an image set with the same model identifier among the multiple images. If there is an image set with the same model identifier among the multiple images, it is determined that there is an image set of the same model among the multiple images; otherwise, it is determined that there is no image set of the same model among the multiple images.
[0081] In this embodiment, images of the same object model in the database are pre-labeled as identical images. Specifically, this labeling can be achieved by assigning a common model identifier to images containing the same object model. For example, if two images both contain a three-seater sofa, and the three-seater sofa in image 1 and image 2 are the same model, then images 1 and 2 can be labeled with the same model identifier. This common model identifier can be a text identifier, such as "three-seater sofa style 1". It can also be an identifier composed of shape symbols or a combination of numbers and letters, and can be set based on actual needs; no limitation is made here. When a user searches for identical images, the common model identifier can be directly used as a query condition. If a set of images with the same model identifier exists among multiple images, it indicates that images of the same three-seater sofa model exist among the multiple images obtained in step 301; otherwise, it is determined that no images of the same three-seater sofa model exist. Identifying identical images through pre-labeled model identifiers eliminates the need for image feature matching in every search, significantly reducing computational load and improving search efficiency compared to deep learning-based methods for determining identical images.
[0082] Step 303: Determine a target image from the set of images of the same model.
[0083] In this step, if there is a set of images of the same model among multiple images, in order to prevent multiple images of the same model from being displayed on one page, a target image can be selected from the set of images of the same model as a representative image for display. In this way, it can ensure that the image of the object model of this style can be displayed and avoid omissions, and it can also avoid the inefficiency caused by displaying multiple images of the same model together, thereby improving the efficiency of users browsing images and improving the user experience.
[0084] In one embodiment, step 303 may specifically include: if a set of images of the same model exists among multiple images, determining whether there is a set of images of the same merchant with the same merchant identifier within the set of images of the same model. If there is a set of images of the same merchant within the set of images of the same model, determining a candidate image of the same model from the set of images of the same merchant. Comparing the usage frequency of the first remaining image of the same model in the set of images of the same model with the candidate image of the same model, determining the image of the same model with the highest usage frequency as the target image.
[0085] In this embodiment, the first remaining identical images refer to images in the identical model image set excluding those from the same merchant's image set. Assume the object model is a model of a product sold by a merchant. For example, in an online home design platform, some home furnishing merchants sell different furniture, and each piece of furniture can be configured with an object model image to assist buyers in viewing relevant information. In this scenario, images belonging to the same object model may belong to different merchants, or multiple identical object model images may belong to the same merchant. For example, the same merchant might upload different images of the same style of three-seater sofa, each image corresponding to an inventory link. That is, the multiple images uploaded by the merchant are simply different inventory items of the same sofa. Since object model images can be linked to merchant links, usually one image is enough to guide users to the corresponding merchant's store to view object models of other products. Therefore, for identical object model images belonging to the same merchant, one candidate image can be selected as a representative image, instead of displaying all of them on the browsing page.
[0086] Specifically, if a set of images of the same three-seater sofa model exists among the multiple images obtained in step 301, the merchant identifiers bound to each image in the set of images of the same model are used to determine whether there is a set of images from the same merchant with the same merchant identifier. If so, it means that multiple images of the same object model belong to the same merchant. A candidate image of the same model is then identified from this set. The usage frequency of the first remaining images of the same model in the set is compared with that of the candidate image, and the image with the highest usage frequency is identified as the target image. Usage frequency refers to how often an image is selected and used by different users on the platform over a period of time, representing the image's popularity. For example, the frequency with which a three-seater sofa model image is selected and applied to home decoration designs by different users within a day is the usage frequency of that three-seater sofa model image. A higher usage frequency indicates that the object model in the image is more popular with users. Comparing the usage frequency of the candidate image with that of the first remaining images of the same model is equivalent to comparing images of the same object model from different merchants, and the image with the highest usage frequency is selected as the target image. This further avoids displaying multiple identical images from the same merchant, saving page resources and improving the efficiency of model image display.
[0087] In one embodiment, step 303 may further include: if there is a set of images from the same merchant in the set of images of the same model, the image with the highest frequency of use in the set of images from the same merchant is determined as a candidate image of the same model.
[0088] In this embodiment, if the image set of the same model contains an image set of the same merchant, the frequency of use can be used as a filtering condition when selecting candidate images of the same model from the image set of the same merchant. The image with the highest frequency of use can be selected as the candidate image of the same model. For example, if the image set of the same merchant includes three images of the same three-seater sofa from merchant A, and the sofas in the three images are red, yellow, and gray respectively, the gray sofa has the highest frequency of use. Therefore, the image of the gray sofa is selected as the candidate image of the same model. In this way, the actual needs of market users are further considered, providing users with more useful information and improving the efficiency of model image display.
[0089] In one embodiment, step 303 may further include: if there is no image set of the same merchant in the image set of the same model, the image of the same model that is used most frequently in the image set of the same model is determined as the target image.
[0090] In this embodiment, if there is a set of images of the same model among the multiple images of the three-seater sofa model obtained in step 301, and the images in the set of images of the same model belong to different merchants, then the images of the same object model from different merchants are compared, and the image with the highest usage frequency is selected as the target image. In this way, even if the images of the same object model belong to different merchants, multiple images of the same model can be avoided from being displayed together, saving page resources and improving the efficiency of model image display.
[0091] Step 304: Display the target image and the untagged non-identical image on the page.
[0092] In this step, the page can be the user's homepage. Non-identical images refer to images that are not marked as identical. Specifically, they refer to images other than the set of identical model images obtained in step 301. The target image and non-identical images constitute a set of model images that are different from each other. These images can be displayed on the homepage. In this way, only different model images are displayed on the homepage, avoiding the inefficiency caused by displaying identical model images multiple times on the same page, improving the image display effect, and enhancing the user experience.
[0093] In one embodiment, step 304 may specifically include: when the total number of the target image and dissimilar images is less than or equal to the image capacity of a page, displaying the target image and dissimilar images on a page in descending order of their usage frequency. When the total number is greater than the image capacity of a page, displaying the target image and dissimilar images on different pages in descending order of their usage frequency.
[0094] In this embodiment, if the final number of images to be displayed does not exceed the image capacity of a single page, they are displayed on the current page in descending order of usage frequency. That is, frequently used images are displayed first to improve image display efficiency. If the final number of images to be displayed exceeds the image capacity of a single page, meaning the current page cannot display all the images, then frequently used images are displayed first on the current page. Once the current page is full, the remaining images are distributed across subsequent pages.
[0095] In one embodiment, if the image set of the same merchant is a set of different color images of the same object model, then after step 304, the method further includes: if there is an image set of the same merchant in the image set of the same model, and the candidate image of the same model is determined as the target image, then the second remaining image of the same model in the image set of the same merchant is configured to be displayed on the next level page of the current page.
[0096] In this embodiment, the second remaining identical image is an image from the same merchant's image set excluding the candidate identical images. The next level page of the current page can be a next level page pointed to by a link on the current page. If the same merchant's image set is a set of images of the same object model in different colors, for example, if the same merchant's image set includes three images of the same three-seater sofa belonging to merchant A, where the sofa in image 1 is red, the sofa in image 2 is yellow, and the sofa in image 3 is gray, when one of the candidate identical images is selected as the target image to be displayed on the current page, assuming that red is used more frequently, and the red sofa model image 1 is selected as the target image, then images 2 and 3 are the corresponding second remaining identical images. The second remaining identical image of this merchant can be displayed in the next level page of the current page where the target image 1 is located. When the user clicks the link on image 1, the corresponding next level page will be displayed, and the second remaining identical image will be displayed in the next level page.
[0097] like Figure 4 The diagram illustrates an example of page display within the user interface 400 of a terminal. The user enters the search command "three-seater sofa" in the search box 401. Image 1 (red) is the target image. Image 1 and other images of the same model are displayed on the current page 410. When the cursor is placed on a specific position on image 1, the corresponding next-level page 411 can be opened. Image 2 and image 3 can be displayed on the next-level page 411. In this way, multiple color schemes of the same object model can be retained, improving the comprehensiveness of the model image display.
[0098] In one embodiment, after step 304, the method may further include: if there is a set of images of the same model among multiple images, and there is no set of images of the same merchant among the set of images of the same model, then the other images of the same model, excluding the target image, are configured to be displayed on different pages.
[0099] In this embodiment, if the collection of images of the same model belongs to different merchants, in order to guide users to different merchants, other images of the same model besides the target image can be displayed on different pages. For example, other images of the same model besides the target image can be displayed on different pages, thus ensuring that multiple images of the same model are not displayed on the same page. Here, there can be multiple different pages, and different pages refer to pages other than the current page. For example, if 100 images of a three-seater sofa are found in step 301, and 10 images are displayed on one page, then 10 pages are needed to display them all. In this case, the first page is the current page, and pages 2 to 10 are the different pages. When displaying other images of the same model on different pages, they can be sorted from high to low according to their usage frequency, displaying the more frequently used images first to improve the user's image browsing experience.
[0100] Step 305: Display multiple images on the page in descending order of their frequency of use.
[0101] In this step, if no set of images of the same model exists among the multiple images, it means that there will be no phenomenon of multiple images of the same model being displayed together. In this case, the multiple images obtained in step 301 can be displayed normally on the page. Specifically, the usage frequency of multiple images can be compared, and the multiple images can be displayed on the page in descending order of usage frequency, showing the more frequently used images first to improve the user's image browsing experience.
[0102] In one embodiment, when displaying a model image, relevant parameter information corresponding to the object model in the image can be displayed on the model image, such as the size information, usage attribute information, and merchant information of the object model, so that users can intuitively understand the product information, improve the efficiency of image display, improve the terminal interaction performance, and improve the user's experience of browsing images.
[0103] The above-described method for displaying model images involves retrieving multiple images of a specific object model when a user searches for it. It then determines if there is a set of images tagged as the same model. If so, a target image is selected from this set and displayed on the page. This ensures that images of the same model are not repeatedly displayed on the same page, avoiding the inefficiency caused by multiple displays of the same object model, improving the display effect of object model images, and enhancing the user's browsing experience.
[0104] Please refer to Figure 5A This is a method for displaying a model image according to an embodiment of this application. The method can be performed by... Figure 1 The electronic device 1 shown is used to perform this action and can be applied to... Figure 2 In the application scenarios of the model image display system shown, this aims to avoid the inefficiency of displaying the same object model image multiple times on the same page, improve the display effect of the object model image, and enhance the user browsing experience. This embodiment takes terminal 220 as the execution end as an example. Compared with the previous embodiment, this embodiment also includes a process of pre-labeling the images in the database as identical. This method includes the following steps:
[0105] Step 501: Obtain multiple raw images of the object model to be processed.
[0106] In this step, the original images can be images of unlabeled object models; that is, each original image can contain an object model. Taking an online home design platform as an example, the original images can be images of home models provided by the platform or images of home models uploaded by merchant users. Original images can be directly read from the platform's database.
[0107] Step 502: Extract the image features of each original image.
[0108] In this step, in order to label the original images with the same type, it is necessary to perform advanced image analysis to extract the image features of each original image. These image features are used to characterize the outline, texture, and other information of the object models in the original images, so as to identify which object models are the same type.
[0109] In one embodiment, step 502 may specifically include: calculating the hash features of each original image; and generating image features for each original image based on the hash features.
[0110] In this step, if we ignore the color of the object model and consider objects with similar or identical outlines and textures as belonging to the same model, we can characterize the image features of each original image by calculating its hash features, and thus characterize the features of the corresponding object model. Hash features can include one or more of the following: average hash value, perceptual hash value, and differential hash value.
[0111] Assuming that the average hash value and perceptual hash value of the original image are used as two hash features to generate the image features of the original image, the specific calculation process for each original image can be as follows:
[0112] 1. Calculate the average hash value aHash of the original images. In real-world scenarios, the result of aHash can be a 64-bit feature vector composed of 0s and 1s. The smaller the Hamming distance between the aHash values of two original images, the higher the similarity between the two original images.
[0113] 2. Calculate the perceptual hash value (pHash) of the original image. In real-world scenarios, the pHash result is a 64-bit feature vector composed of 0s and 1s. The smaller the Hamming distance between the pHash values of two original images, the higher the similarity between the two original images.
[0114] 3. Merge aHash and pHash as image features of the original image. At this point, the image features of the original image are 128-bit feature vectors.
[0115] The specific process of aHash is illustrated below:
[0116] (1) Read the original image.
[0117] (2) Convert the original image into a grayscale image.
[0118] (3) Scale the grayscale image to 8x8 to ignore the error caused by the slight displacement of the image. The interpolation method can be Antialias (an anti-aliasing interpolation method).
[0119] (4) Calculate the average pixel value of the scaled grayscale image.
[0120] (5) Determine the relationship between the pixel value of the scaled grayscale image and the average pixel value of the scaled grayscale image. If the value of the current pixel in the scaled grayscale image is greater than the average pixel value of the scaled grayscale image, then the feature value of the current pixel is 1; otherwise, the feature value of the current pixel is 0.
[0121] (6) By making the pixel values of the scaled grayscale image more thick in the manner described in (5), a 64-bit aHash feature vector consisting of 0 and 1 can be obtained. The aHash feature vector can highlight the contour features of the object model in the original image.
[0122] The specific process of pHash is illustrated below:
[0123] (1) Read the original image.
[0124] (2) Convert the original image into a grayscale image.
[0125] (3) Scale the grayscale image to a size of 32x32, and the interpolation method can be Antialias.
[0126] (4) Perform Discrete Cosine Transform (DCT) on the scaled grayscale image.
[0127] (5) The upper left 8x8 matrix region is taken as the low-frequency DCT feature. The low-frequency DCT feature can represent the main features of the object model in the original image. This is different from the commonly used pHash calculation method. The upper left 8x8 matrix region generally contains the features of the main part of the object model. By selectively taking the upper left 8x8 matrix region as the low-frequency DCT feature, compared with the 32x32 DCT feature in the commonly used method, it can better highlight the main body of the object model, so as to more accurately distinguish the main body of the object model based on the pHash feature vector.
[0128] (6) Calculate the median of low-frequency DCT features.
[0129] (7) Determine the relationship between the DCT feature and the median of the DCT feature at each position of the scaled grayscale image. If the DCT feature value at the current position is greater than the median of the DCT feature, then assign the value 1 to the current position; otherwise, assign the value 0.
[0130] (8) Thus, a 64-bit pHash feature vector consisting of 0 and 1 is obtained. The pHash feature vector can highlight the main features of the object model in the original image.
[0131] In one embodiment, the difference hash value dhash is more sensitive to mutation information in an image. For an original image, the corresponding dhash feature vector, aHash feature vector and pHash feature vector can be concatenated together as the image feature of the original image, or the dhash feature vector can be used as the image feature of the original image alone. The combination method can be flexibly set according to actual needs, and this embodiment does not limit it.
[0132] In one embodiment, step 502 may further include: calculating the color mean of each channel in each original image, and generating image features of each original image based on the color mean and hash features.
[0133] In this embodiment, strict matching judgment can also consider color. The color mean can characterize the color performance of an image and can be used as the color feature of the original image. The color feature can be concatenated with the hash feature to form the image feature of the original image, thus increasing color discrimination and improving the discrimination precision of the matching object model.
[0134] Assuming that the image features of the original image are generated using three features—the color mean, the average hash value, and the perceptual hash value—the specific calculation process for each original image is as follows: Figure 5B As shown, it may include the following steps:
[0135] S1. Calculate the mean color value of the three channels R (red), G (green), and B (blue) of the original image, which ranges from 0 to 255.
[0136] S2. Pad the color mean of each channel with 0 to 3 bits to generate a 9-bit color feature vector. In this way, the color feature can occupy the first 9 bits in the image features of the original image.
[0137] S3. Calculate the average hash value aHash of the original images. In real-world scenarios, the result of aHash can be a 64-bit feature vector composed of 0s and 1s. The smaller the Hamming distance between the aHash values of two original images, the higher the similarity between the two original images.
[0138] S4. Calculate the perceptual hash value (pHash) of the original images. In real-world scenarios, the pHash result is a 64-bit feature vector composed of 0s and 1s. The smaller the Hamming distance between the pHash values of two original images, the higher the similarity between the two original images.
[0139] S5. Combine the color mean, aHash, and pHash as image features of the original image. At this point, the image features of the original image are 137-bit feature vectors.
[0140] Step 503: Mark two original images whose image features have a similarity greater than a preset threshold as the same.
[0141] In this step, a preset threshold is used to determine whether two original images belong to the same object model. Only two original images with sufficiently high similarity will be marked as the same. The preset threshold can be set based on actual needs. When a higher accuracy of matching is required, the preset threshold can be increased. Assuming the similarity value ranges from (0, 1), a similarity of 1 indicates that the two original images are identical. If the requirement in practice is that the images must be identical, then the preset threshold can be set to 1.
[0142] In one embodiment, two original images marked as identical can be tagged with a "same style" identifier and stored in a database. Specifically, this tagging can be done by assigning a "same model" identifier to original images containing the same object model. For example, if two images both contain a three-seater sofa, and the three-seater sofa in image 1 and image 2 are the same style, then image 1 and image 2 can be tagged with "same model". This "same model" identifier can be a text identifier, such as "three-seater sofa style 1". It can also be an identifier composed of shape symbols or a combination of numbers and letters, and can be flexibly set based on actual needs; no limitation is imposed here.
[0143] In this embodiment, calculating image features requires only a limited number of matrix calculations, achieving a faster feature vector construction speed than deep learning methods. Compared to the large number of matrix calculations involved in deep learning methods, this greatly simplifies the amount of data computation.
[0144] Step 504: In response to the user's search command, acquire multiple images matching the search command, each image including an object model matching the search command. See the detailed description of step 301 in the above embodiments.
[0145] Step 505: Determine whether there exists a set of images with the same model identifier among the multiple images. If yes, proceed to step 506; otherwise, proceed to step 511.
[0146] Step 506: If multiple images contain a set of images with the same model identifier, determine that multiple images contain a set of images with the same model. See the detailed description of step 302 in the above embodiments.
[0147] Step 507: If multiple images contain a set of images of the same model, determine whether there is a set of images of the same model from the same merchant with the same merchant identifier. If yes, proceed to step 508; otherwise, proceed to step 510. See the detailed description of step 303 in the above embodiments.
[0148] Step 508: If the set of images of the same model contains an image set of the same merchant, determine a candidate image of the same model from the image set of the same merchant. See the detailed description of step 303 in the above embodiments.
[0149] Step 509: Compare the usage frequency of the first remaining identical images in the identical model image set with that of the candidate identical images. The identical image with the highest usage frequency is determined as the target image. The first remaining identical images are images in the identical model image set excluding those from the same merchant's image set. Then proceed to step 511. See the detailed description of step 303 in the above embodiments.
[0150] Step 510: If the image set of the same model does not contain an image set of the same merchant, the image of the same model with the highest frequency of use in the image set of the same model is determined as the target image. Then proceed to step 511. See the relevant description of step 303 in the above embodiments for details.
[0151] Step 511: Display the target image and the unmarked non-identical image on the page. See the detailed description of step 304 in the above embodiments.
[0152] Step 512: Determine that there is no set of images with the same model among multiple images, and display the multiple images on the page in descending order of their usage frequency. See the detailed description of step 305 in the above embodiments.
[0153] The above-described method for displaying model images generates an integer feature vector for each original image, which is simpler to calculate than the floating-point feature vector obtained by traditional deep learning methods. This integer feature vector can be stored in a database. When a user needs to determine if there are similar images, they only need to search the database index. Compared with traditional deep learning methods, it is not necessary to calculate the similarity between the feature vector of the searched image and the feature vectors of all model images in the library for each retrieval, which greatly reduces the computational complexity.
[0154] On the other hand, the embodiments of this application can satisfy customized same-item judgment. The image features obtained based on the color mean and image hash features can be stored in a relational database, which can better support same-item judgment across different dimensions. Traditional deep learning methods require indexing global feature vectors, resulting in search engines finding only global same-items and failing to perform customized differentiation, such as distinguishing same-items only at the merchant level. Therefore, the solution of the embodiments of this application can improve the customization and flexibility of same-item judgment.
[0155] Please refer to Figure 6 This is a method for recognizing images of the same model according to an embodiment of this application. The method can be performed by... Figure 1 The electronic device 1 shown is used to perform this action and can be applied to... Figure 2 In the application scenario of the model image display system shown, the goal is to improve the efficiency of recognizing images of the same model. This embodiment takes terminal 220 as the execution end as an example, and the method includes the following steps:
[0156] Step 601: Obtain multiple images to be processed, each image containing an object model with the same attributes.
[0157] In this step, "same attributes" means that the object models in multiple images to be processed belong to the same category, such as all being chair models or all being sofa models.
[0158] Step 602: Determine the model identifier corresponding to the object model in each image to be processed.
[0159] Step 603: Determine whether there exists an image set with the same model identifier among the multiple images to be processed. If yes, proceed to step 604; otherwise, it means that there are no images of the same model among the multiple images to be processed.
[0160] Step 604: If there is a set of images with the same model identifier among multiple images to be processed, determine that the image set is the same type of image.
[0161] The aforementioned method for identifying identical model images does not require comparing feature vectors between images; it only needs to compare the model identifiers of the images to be processed to determine which images belong to the same model. This significantly improves data processing speed and saves data resources. Please refer to the relevant descriptions in the above embodiments for details. Further elaboration is omitted here.
[0162] Please refer to Figure 7 This is a method for displaying a model image according to an embodiment of this application. The method can be performed by... Figure 1 The electronic device 1 shown is used to perform this action and can be applied to... Figure 2 In the application scenario of the model image display system shown, this method aims to avoid the inefficiency of displaying the same object model image multiple times on the same page, improve the display effect of the object model image, and enhance the user browsing experience. This embodiment takes terminal 220 as the execution end as an example, and takes the product image display scenario of an online home design platform as an example. The method includes the following steps:
[0163] Step 701: In response to the user's search command for a home furnishing model, acquire multiple images that match the search command, each image including a home furnishing model that matches the search command.
[0164] Step 702: Determine if a set of images with the same model exists among multiple images. The furniture models in each image of the set of images with the same model are marked as the same. If yes, proceed to step 703; otherwise, proceed to step 705.
[0165] Step 703: If there is a set of images of the same model among multiple images, determine a target image from the set of images of the same model.
[0166] Step 704: Display the target image and non-identical images on the page. Non-identical images are images other than the set of images of the same model from among multiple images.
[0167] Step 705: Display multiple images on the page in descending order of their frequency of use.
[0168] The above-described method for displaying model images involves retrieving multiple images of a product model when a user searches for it. It then determines if there is a set of images marked as identical to that model. If so, a target image is selected from this set and displayed on the page along with any images of the same model. This ensures that images of identical models are not repeatedly displayed on the same page, avoiding the inefficiency caused by multiple displays of identical product models, improving the display effect of product model images, and enhancing the user browsing experience. For details, please refer to the relevant descriptions in the above embodiments. Further details are omitted here.
[0169] Please refer to Figure 8 This is a model image display device 800 according to an embodiment of this application. The device can be generated by... Figure 1 The electronic device 1 shown is used to perform this action and can be applied to... Figure 2 In the application scenarios shown, the display system for model images aims to avoid the inefficiency of displaying the same object model image multiple times on the same page, improve the display effect of object model images, and enhance the user browsing experience. The device includes:
[0170] The first acquisition module 801 is used to acquire multiple images that match the user's search command in response to the user's search command, and each image includes an object model that matches the search command.
[0171] The judgment module 802 is used to determine whether there is a set of images with the same model among multiple images. The object models of each image in the set of images with the same model are marked as the same.
[0172] The determination module 803 is used to determine a target image from the set of images of the same model if there is a set of images of the same model among multiple images.
[0173] Display module 804 is used to display the target image and non-identical images on the page. Non-identical images are images other than the set of images of the same model from multiple images.
[0174] In one embodiment, the determination module 802 is used to determine whether there is an image set with the same model identifier among multiple images. If there is an image set with the same model identifier among multiple images, it is determined that there is an image set with the same model among multiple images; otherwise, it is determined that there is no image set with the same model among multiple images.
[0175] In one embodiment, the determining module 803 is used to determine whether, if a set of images of the same model exists among multiple images, there is a set of images of the same merchant with the same merchant identifier. If a set of images of the same merchant exists among the set of images of the same model, a candidate image of the same model is determined from the set of images of the same merchant. The usage frequency of the first remaining images of the same model in the set of images of the same model is compared with that of the candidate images of the same model, and the image of the same model with the highest usage frequency is determined as the target image. The first remaining images of the same model are the images in the set of images of the same model excluding the set of images of the same merchant.
[0176] In one embodiment, the determining module 803 is used to determine the image with the highest frequency of use in the image set of the same model as a candidate image of the same model if there is an image set of the same merchant in the image set of the same merchant.
[0177] In one embodiment, the same merchant image set is a set of images of the same object model in different colors. After displaying the target image and the non-same-model images on the page, the display module 804 is further configured to, if the same merchant image set exists in the same model image set and the candidate same-model image is determined as the target image, configure the second remaining same-model image in the same merchant image set to be displayed on the next level page of the current page. The second remaining same-model image is the image in the same merchant image set excluding the candidate same-model image.
[0178] In one embodiment, the determining module 803 is used to determine the most frequently used image of the same model in the same model image set as the target image if there is no image set of the same merchant in the same model image set.
[0179] In one embodiment, if there is a set of images of the same model among multiple images, after displaying the target image and the images of different models on the page, the display module 804 is further configured to configure the third remaining images of the same model in the set of images of the same model on different pages if there is no set of images of the same merchant in the set of images of the same model. The third remaining images of the same model are images of the same model in the set of images other than the target image on different pages.
[0180] In one embodiment, the system further includes: a second acquisition module, configured to acquire multiple original images of the object model to be processed before acquiring multiple images matching the search command in response to the user's search command; an extraction module, configured to extract image features from each original image; and a labeling module, configured to label two original images whose image feature similarity is greater than a preset threshold as the same type.
[0181] In one embodiment, the extraction module is used to calculate the hash features of each original image, the hash features including one or more of the average hash value, perceptual hash value, and differential hash value. Based on the hash features, image features of each original image are generated.
[0182] In one embodiment, the extraction module is further configured to calculate the color mean of each channel in each original image, and generate image features of each original image based on the color mean and hash features.
[0183] For a detailed description of the display device 800 for the above model image, please refer to the description of the relevant method steps in the above embodiment. The implementation principle and technical effect are similar, and will not be repeated here in this embodiment.
[0184] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the method of any of the foregoing embodiments.
[0185] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the method of any of the foregoing embodiments.
[0186] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of modules is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed.
[0187] The integrated modules described above, implemented as software functional modules, can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods of the various embodiments of this application.
[0188] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the application can be directly manifested as being executed by a hardware processor, or executed by a combination of hardware and software modules within the processor. The memory may include high-speed RAM, and may also include non-volatile memory (NVM), such as at least one disk storage device, and may also be a USB flash drive, external hard drive, read-only memory, disk, or optical disc, etc.
[0189] The aforementioned storage medium can be implemented from 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. The storage medium can be any available medium accessible to general-purpose or special-purpose computers.
[0190] An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Alternatively, the storage medium can be an integral part of the processor. Both the processor and the storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and storage medium can exist as discrete components in an electronic device or host device.
[0191] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.
[0192] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.
[0193] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. 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 is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods of the various embodiments of this application.
[0194] The collection, storage, use, processing, transmission, provision, and disclosure of user data and other information involved in the technical solution of this application all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.
[0195] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.
Claims
1. A display method of a model image, characterized by, The method includes: Obtain multiple raw images of the object model to be processed; Calculate the hash feature of each of the original images and the average color value of each channel in each of the original images respectively; Based on the color mean and the hash feature, image features of each of the original images are generated; Two original images whose image features have a similarity greater than a preset threshold are labeled as having the same design. In response to a user's search command, multiple images matching the search command are obtained, each image including an object model matching the search command; Determine whether there exists a set of images with the same model identifier among the multiple images; If there is a set of images with the same model identifier among the multiple images, then it is determined that there is a set of images with the same model among the multiple images, and a target image is determined from the set of images with the same model. The target image is the image with the highest usage frequency in the set of images with the same model. The object models of each image in the set of images with the same model are pre-assigned with the same model identifier to mark them as the same model. The target image and non-identical images are displayed on the page. The non-identical images are images other than the set of identical model images from the plurality of images.
2. The method of claim 1, wherein, Also includes: If there is no set of images with the same model identifier among the plurality of images, then it is determined that there is no set of images with the same model among the plurality of images.
3. The method of claim 1, wherein, Determining a target image from the set of images of the same model includes: Determine whether there exists a set of images of the same model with the same merchant identifier within the set of images of the same model; If the set of images of the same model exists in the set of images of the same merchant, then a candidate image of the same model is determined from the set of images of the same merchant; By comparing the usage frequency of the first remaining identical images in the same model image set with that of the candidate identical images, the identical image with the highest usage frequency is determined as the target image. The first remaining identical images are images in the same model image set excluding the images from the same merchant image set.
4. The method of claim 3, wherein, If the set of images of the same model exists within the set of images of the same merchant, determining a candidate image of the same model from the set of images of the same merchant includes: If the set of images of the same model exists within the set of images of the same merchant, the image with the highest frequency of use in the set of images of the same merchant is determined as the candidate image of the same model.
5. The method of claim 3, wherein, The set of images from the same merchant is a set of images of the same object model in different colors; after displaying the target image and the images of different models on the page, the method further includes: If the image set of the same model exists in the image set of the same merchant, and the candidate image of the same model is determined as the target image, the second remaining image of the same model in the image set of the same merchant is configured to be displayed on the next level page of the current page. The second remaining image of the same model is the image in the image set of the same merchant excluding the candidate image of the same model.
6. The method of claim 3, wherein, If a set of images of the same model exists among the plurality of images, determining a target image from the set of images of the same model further includes: If the image set of the same model does not exist in the image set of the same merchant, the image of the same model that is used most frequently in the image set of the same model is determined as the target image.
7. The method of claim 3, wherein, If the set of images of the same model exists among the multiple images, after displaying the target image and the non-identical images on the page, the method further includes: If the image set of the same model does not exist in the image set of the same merchant, the third remaining image of the same model in the image set of the same model will be configured to be displayed on a different page. The third remaining image of the same model is a different page from the images in the image set of the same model excluding the target image.
8. A display method of a model image, characterized by, include: Obtain multiple raw images of the home furnishing model to be processed; Calculate the hash feature of each of the original images and the average color value of each channel in each of the original images respectively; Based on the color mean and the hash feature, image features of each of the original images are generated; Two original images whose image features have a similarity greater than a preset threshold are labeled as having the same design. In response to a user's search command for a home furnishing model, multiple images matching the search command are obtained, each image including a home furnishing model matching the search command; Determine whether there exists a set of images with the same model identifier among the multiple images; If there is a set of images with the same model identifier among the multiple images, then it is determined that there is a set of images with the same model among the multiple images, and a target image is determined from the set of images with the same model. The target image is the image with the highest frequency of use in the set of images with the same model, and the home furnishing models of each image in the set of images with the same model are marked as the same. The target image and non-identical images are displayed on the page. The non-identical images are images other than the set of identical model images from the plurality of images.
9. An electronic device, comprising: include: At least one processor; as well as A memory that is communicatively connected to the at least one processor; The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, cause the electronic device to perform the method according to any one of claims 1-8.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, implement the method as described in any one of claims 1-8.
11. A computer program product comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-8.
Citation Information
Patent Citations
Commodity recommendation method and device based on user behavior data
CN113837842A
Same-type commodity aggregation method and device, equipment, medium and product
CN113962773A
Intelligent clothes recommendation method
CN114219556A
