Method and device for generating cabin personalized theme, electronic equipment and storage medium

Abstract

The application discloses a cockpit personalized theme generation method and device, electronic equipment and a storage medium, and the method comprises the following steps: generating a target style cockpit wallpaper according to a user instruction; splicing a plurality of original application icons into an icon matrix; extracting a style feature vector of the wallpaper by using an IP-Adapter, and acquiring structure control information of the icon matrix; inputting the style feature vector and the structure control information into a diffusion model as a composite condition to generate a stylized icon matrix with unified style and maintained structure; and finally, cutting the stylized icon matrix into each application mutation and applying it to the corresponding application program. The application solves the problem of interface element style fragmentation in the prior art, realizes cockpit theme personalized generation, and improves visual uniformity and user experience.

Description

Technical Field

[0001] This invention relates to the field of automotive and artificial intelligence intersection technology, and in particular to methods and apparatus for generating personalized cockpit themes, electronic devices, and storage media. Background Technology

[0002] With the increasing demand for personalized smart cockpits, the wallpapers displayed on the infotainment systems of mainstream automakers are generally pre-stored by users. While switching wallpapers using these pre-stored wallpapers provides a fast and stable user experience, its inherent disadvantages cannot be ignored. The number and variety of predefined wallpapers in the current settings are limited, making it difficult to meet the increasingly diverse and personalized needs of users. As users' requirements for the aesthetics and personalization of the in-car environment increase, single and fixed wallpaper and icon options appear too restrictive, lacking novelty and appeal. Although predefined wallpapers allow for quick switching, this also means that these wallpapers lack dynamic changes and the possibility of personalized customization, resulting in a relatively monotonous experience. Furthermore, the independent generation of wallpapers and icons leads to stylistic fragmentation and inefficient manual adjustments.

[0003] Current in-vehicle infotainment systems generally use pre-stored static image libraries for wallpaper customization, which has significant limitations: the fixed number and styles of wallpaper resources result in monotonous and unattractive content, failing to meet users' dynamic aesthetic needs; although some solutions attempt to introduce AI-generated content (AIGC) technology to improve wallpaper quality, they only focus on the single area of ​​wallpaper generation and have not yet solved the problem of visual coordination between wallpapers and system icons, resulting in a disjointed style of interface elements, requiring users to make tedious manual adjustments. Therefore, the cockpit urgently needs an efficient generation solution that can achieve a unified style for multiple cockpit elements through natural interaction. Summary of the Invention

[0004] The purpose of this invention is to address the problem that the existing single and fixed wallpaper and icon options for customers are too limited and lack novelty and appeal. It provides a method and device for generating personalized cockpit themes, electronic devices, and storage media. It achieves a unified style of vehicle interface elements through an end-to-end personalized icon generation architecture and improves the icon generation speed through icon matrix processing technology.

[0005] The technical solution adopted in this invention is:

[0006] A method for generating personalized cockpit themes, applied to a vehicle-cloud collaborative system, includes the following:

[0007] Generates a target style cockpit wallpaper corresponding to the user's instructions;

[0008] Obtain multiple original application icons within the cockpit system and concatenate them into an icon matrix.

[0009] Extracting style feature vectors from target style cockpit wallpapers;

[0010] Obtain the structural control information of the icon matrix;

[0011] The text description corresponding to the user instruction is used as the first generation condition, the style feature vector is used as the second generation condition, and the structural control information is used as the third generation condition. These are all input into the diffusion model to generate a stylized icon matrix that is consistent with the target style cockpit wallpaper style and maintains the original structure.

[0012] The stylized icon matrix is ​​divided into multiple independent stylized application icons;

[0013] The target style cockpit wallpaper and multiple stylized application icons were distributed and applied to the cockpit display interface.

[0014] Furthermore, the implementation of generating a target-style cockpit wallpaper corresponding to the user's instructions includes:

[0015] Receive user commands in voice format and convert them into text commands;

[0016] Perform security audits on text commands;

[0017] After approval, the text instructions are enhanced using a large language model to obtain enhanced prompts.

[0018] The enhanced prompt words are input into the text-based image model to generate the target style cockpit wallpaper.

[0019] Furthermore, after generating the target style cockpit wallpaper, it also includes:

[0020] Perform image content security review on the target style cockpit wallpaper;

[0021] After approval, the target style cockpit wallpaper will be stored on a cloud server.

[0022] Furthermore, multiple original application icons are concatenated into an icon matrix, including:

[0023] Multiple original application icons with preset resolutions are arranged and stitched together according to an N*M grid to generate a combined matrix image, and the position coordinate information of each original application icon in the matrix image is recorded.

[0024] Furthermore, the implementation of extracting style feature vectors based on target style cockpit wallpapers includes:

[0025] The target style cockpit wallpaper is encoded using the IP-Adapter image feature adapter to obtain a style feature vector, which is then used to inject style information into the denoising process of the diffusion model.

[0026] Furthermore, the implementation of obtaining the structural control information of the icon matrix includes:

[0027] Edge detection is performed on the icon matrix to obtain the outline structure map of all original application icons in the icon matrix; the structural control information is the outline structure map.

[0028] Furthermore, the implementation of inputting structural control information as a third generation condition into the diffusion model includes:

[0029] The ControlNet control network is used to guide the generation process of the diffusion model by using the contour structure map as a control condition, so that the contour structure of each icon in the generated stylized icon matrix is ​​consistent with the corresponding original application icon.

[0030] Furthermore, the implementation of dividing the stylized icon matrix into multiple independent stylized application icons includes: dividing the stylized icon matrix into independent icons corresponding to the original number of application icons based on the position coordinate information.

[0031] Furthermore, after dividing the stylized icon matrix into multiple independent stylized application icons, the process also includes: performing image content security audits on each of the divided stylized application icons.

[0032] A device for generating personalized cockpit themes, applied to a vehicle-cloud collaborative system, includes:

[0033] The wallpaper generation module is used to generate cockpit wallpapers in the target style corresponding to user commands.

[0034] The matrix building module is used to obtain multiple original application icons within the cockpit system and concatenate these icons into an icon matrix.

[0035] The feature extraction module is used to extract style feature vectors based on the target style cockpit wallpaper;

[0036] The structure acquisition module is used to acquire the structural control information of the icon matrix;

[0037] The collaborative generation module is used to input the text description corresponding to the user's instruction as the first generation condition, the style feature vector as the second generation condition, and the structural control information as the third generation condition into the diffusion model to generate a stylized icon matrix that is consistent with the target style cockpit wallpaper style and maintains the original structure.

[0038] The icon segmentation module is used to segment the stylized icon matrix into multiple independent stylized application icons;

[0039] The theme application module is used to distribute and apply target-style cockpit wallpapers and multiple stylized application icons to the cockpit display interface.

[0040] The present invention also discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps of the method for generating the personalized cockpit theme.

[0041] The present invention also discloses a computer-readable storage medium storing a computer program thereon, wherein the computer program, when executed by a processor, implements the steps of the method for generating the personalized cockpit theme.

[0042] The present invention adopts the above technical solution and has the following beneficial effects compared with the prior art:

[0043] This invention, through an architecture based on IP-Adapter, ControlNet, and a diffusion model, uses the generated personalized wallpaper as a style blueprint to drive a batch of icons to undergo a deep stylized redraw while accurately maintaining the original functional structure. This solves the problem of the disconnect between wallpaper and icon styles in existing solutions and achieves a highly harmonious and unified visual theme from wallpaper to system icons.

[0044] This invention employs icon matrix processing technology to improve upon the traditional method of generating icons one by one, transforming it into batch generation. This significantly reduces the number of model calls and overall computation time, shortening the waiting period from when a user issues a command to when they receive the complete theme. Users only need a single natural language description to trigger the synchronous generation of "wallpaper-icons," simplifying the originally tedious process of manual selection, matching, and adjustment into a fully automated workflow, greatly improving the efficiency and convenience of personalized customization.

[0045] This invention leverages ControlNet's structural control capabilities to ensure that stylized icons remain clearly legible, avoiding functional recognition obstacles caused by excessive artistic processing, and guaranteeing interaction efficiency and safety in driving scenarios. Furthermore, multiple content security review mechanisms can be selected to effectively filter inappropriate information, ensuring the compliance and reliability of generated content, enabling the technology to meet the stringent requirements of the in-vehicle environment while pursuing personalization. Attached Figure Description

[0046] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments;

[0047] Figure 1 This is a flowchart illustrating the method for generating personalized cockpit themes according to the present invention. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0049] like Figure 1 As shown, this invention discloses a method for generating personalized cockpit themes, applied to a vehicle-cloud collaborative system. The method includes:

[0050] Generates a target style cockpit wallpaper corresponding to the user's instructions;

[0051] Obtain multiple original application icons within the cockpit system and concatenate them into an icon matrix.

[0052] Extracting style feature vectors from target style cockpit wallpapers;

[0053] Obtain the structural control information of the icon matrix;

[0054] The text description corresponding to the user instruction is used as the first generation condition, the style feature vector is used as the second generation condition, and the structural control information is used as the third generation condition. These are all input into the diffusion model to generate a stylized icon matrix that is consistent with the target style cockpit wallpaper style and maintains the original structure.

[0055] The stylized icon matrix is ​​divided into multiple independent stylized application icons;

[0056] The target style cockpit wallpaper and multiple stylized application icons were distributed and applied to the cockpit display interface.

[0057] Furthermore, the implementation of generating a target-style cockpit wallpaper corresponding to the user's instructions includes:

[0058] Receive user commands in voice format and convert them into text commands;

[0059] Perform security audits on text commands;

[0060] After approval, the text instructions are enhanced using a large language model to obtain enhanced prompts.

[0061] The enhanced prompt words are input into the text-based image model to generate the target style cockpit wallpaper.

[0062] Furthermore, after generating the target style cockpit wallpaper, it also includes:

[0063] Perform image content security review on the target style cockpit wallpaper;

[0064] After approval, the target style cockpit wallpaper will be stored on a cloud server.

[0065] Furthermore, multiple original application icons are concatenated into an icon matrix, including:

[0066] Multiple original application icons with preset resolutions are arranged and stitched together according to an N*M grid to generate a combined matrix image, and the position coordinate information of each original application icon in the matrix image is recorded.

[0067] Furthermore, the implementation of extracting style feature vectors based on target style cockpit wallpapers includes:

[0068] The target style cockpit wallpaper is encoded using the IP-Adapter image feature adapter to obtain a style feature vector, which is then used to inject style information into the denoising process of the diffusion model.

[0069] Furthermore, the implementation of obtaining the structural control information of the icon matrix includes:

[0070] Edge detection is performed on the icon matrix to obtain the outline structure map of all original application icons in the icon matrix; the structural control information is the outline structure map.

[0071] Furthermore, the implementation of inputting structural control information as a third generation condition into the diffusion model includes:

[0072] The ControlNet control network is used to guide the generation process of the diffusion model by using the contour structure map as a control condition, so that the contour structure of each icon in the generated stylized icon matrix is ​​consistent with the corresponding original application icon.

[0073] Furthermore, the implementation of dividing the stylized icon matrix into multiple independent stylized application icons includes: dividing the stylized icon matrix into independent icons corresponding to the original number of application icons based on the position coordinate information.

[0074] Furthermore, after dividing the stylized icon matrix into multiple independent stylized application icons, the process also includes: performing image content security audits on each of the divided stylized application icons.

[0075] A device for generating personalized cockpit themes, applied to a vehicle-cloud collaborative system, includes:

[0076] The wallpaper generation module is used to generate cockpit wallpapers in the target style corresponding to user commands.

[0077] The matrix building module is used to obtain multiple original application icons within the cockpit system and concatenate these icons into an icon matrix.

[0078] The feature extraction module is used to extract style feature vectors based on the target style cockpit wallpaper;

[0079] The structure acquisition module is used to acquire the structural control information of the icon matrix;

[0080] The collaborative generation module is used to input the text description corresponding to the user's instruction as the first generation condition, the style feature vector as the second generation condition, and the structural control information as the third generation condition into the diffusion model to generate a stylized icon matrix that is consistent with the target style cockpit wallpaper style and maintains the original structure.

[0081] The icon segmentation module is used to segment the stylized icon matrix into multiple independent stylized application icons;

[0082] The theme application module is used to distribute and apply target-style cockpit wallpapers and multiple stylized application icons to the cockpit display interface.

[0083] The present invention also discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps of the method for generating the personalized cockpit theme.

[0084] The present invention also discloses a computer-readable storage medium storing a computer program thereon, wherein the computer program, when executed by a processor, implements the steps of the method for generating the personalized cockpit theme.

[0085] The working principle of this invention will be explained in detail below through specific examples:

[0086] The method for generating personalized cockpit themes in this specific implementation includes the following steps:

[0087] S01: Voice Recognition and Security Audit

[0088] S01-1: The vehicle-mounted voice module collects user commands (such as "generate a neon-style urban night scene theme") and converts the voice commands into text using ASR technology;

[0089] S01-2: Upload the instruction text and vehicle-mounted APP information (icons) to the cloud-based large model;

[0090] S01-3: The text security audit engine filters sensitive content. If sensitive content elements (pornography, violence, etc.) are detected, the system terminates the generation process and issues a relevant prompt.

[0091] S02: Enhanced prompts and text-based image generation.

[0092] S02-1: The large language model parses the semantics of instructions and enhances prompt words (e.g., neon-lit cityscape at night, futuristic cyberpunk style, extreme prosperity, dense and towering skyscrapers covered with huge dynamic neon billboards and holographic projections, shimmering with light. Streets flow with liquid rainbow-like light, and the wet asphalt pavement perfectly reflects the colorful neon lights. Strong color contrast: saturated magenta, electric blue, bright yellow, and neon green intertwine and collide. A thin mist and drizzle permeate the air, and the light scatters and diffuses in the air, creating a psychedelic atmosphere. The building surfaces are covered with luminous lines and geometric patterns, rich in detail: rusty pipes, steam overflowing from vents, and neon lights intertwine. Ultra-wide-angle low-view composition, exaggerated perspective enhances the sense of oppression and immersion, and the depth of field blur highlights the light spots. Style: high detail, high sharpness, high contrast, 8K resolution, Unreal Engine rendering, studio lighting, volumetric lighting effects, and light pollution aesthetics.).

[0093] S02-2: Using the large model of text-based images, generate the desired wallpaper image;

[0094] Input: Enhanced prompt words;

[0095] Output: Image;

[0096] S02-3: Image Security Review

[0097] The system uses a pornography and violence detection model to check for pornographic or violent elements in images. If any are found, the generation process is interrupted, and a message appears indicating that the image generation failed and requesting a retry.

[0098] S02-4: Stored on the image server.

[0099] S03: Based on the style of the generated images, batch-produce application icons that match the style.

[0100] S03-1: Icon Matrix Construction

[0101] Arrange the icons on the vehicle's infotainment system (192*192 resolution) into a 4x4 layout to create a combined image (768*768 resolution). Preserve the coordinate mapping relationship to prepare for subsequent icon segmentation.

[0102] S03-2: Perform Canny boundary extraction on the icons in the icon matrix to obtain the structural information of the original icons.

[0103] S03-3: The generated wallpaper is processed using the IP-Adapter model. The instructions for generating the in-vehicle wallpaper and icons are converted into feature vectors, which are then incorporated into the generated features during the subsequent diffusion model denoising process.

[0104] S03-4: Use the ControlNet model to take the structural information of the original icon (the boundary structure processed by Canny) as a condition to guide the generation of the diffusion model.

[0105] S03-5: Based on wallpaper features, original icon features, and prompt words as inputs to the diffusion model, icon generation is performed.

[0106] S03-6: Accurate Icon Reproduction

[0107] Based on a preset coordinate mapping relationship, the stylized composite image is divided into independent icons.

[0108] S03-7: Utilize a pornography and violence detection model to review images for pornography, violence, or other offensive elements. If such elements are found, the request is terminated.

[0109] S04: Departure Machine under Wallpaper and Icons

[0110] S04-1: The wallpaper and stylized icon package are distributed to the display unit via the vehicle bus; and the user is notified via TTS technology that the personalized theme has been generated and is requested to view it.

[0111] S04-2: After user confirmation, the vehicle terminal replaces the generated wallpaper and stylized icons.

[0112] The present invention adopts the above technical solution and has the following beneficial effects compared with the prior art:

[0113] This invention, through an architecture based on IP-Adapter, ControlNet, and a diffusion model, uses the generated personalized wallpaper as a style blueprint to drive a batch of icons to undergo a deep stylized redraw while accurately maintaining the original functional structure. This solves the problem of the disconnect between wallpaper and icon styles in existing solutions and achieves a highly harmonious and unified visual theme from wallpaper to system icons.

[0114] This invention employs icon matrix processing technology to improve upon the traditional method of generating icons one by one, transforming it into batch generation. This significantly reduces the number of model calls and overall computation time, shortening the waiting period from when a user issues a command to when they receive the complete theme. Users only need a single natural language description to trigger the synchronous generation of "wallpaper-icons," simplifying the originally tedious process of manual selection, matching, and adjustment into a fully automated workflow, greatly improving the efficiency and convenience of personalized customization.

[0115] This invention leverages ControlNet's structural control capabilities to ensure that stylized icons remain clearly legible, avoiding functional recognition obstacles caused by excessive artistic processing, and guaranteeing interaction efficiency and safety in driving scenarios. Furthermore, multiple content security review mechanisms can be selected to effectively filter inappropriate information, ensuring the compliance and reliability of generated content, enabling the technology to meet the stringent requirements of the in-vehicle environment while pursuing personalization.

[0116] Obviously, the described embodiments are only a part of the embodiments of this application, not all of them. Without conflict, the embodiments and features in the embodiments of this application can be combined with each other. The components of the embodiments of this application described and illustrated herein can generally be arranged and designed in various different configurations. Therefore, the detailed description of the embodiments of this application is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

Claims

1. A method for generating personalized cockpit themes, applied to a vehicle-cloud collaborative system, characterized in that, The methods include: Generates a target style cockpit wallpaper corresponding to the user's instructions; Obtain multiple original application icons within the cockpit system and concatenate them into an icon matrix. Extracting style feature vectors from target style cockpit wallpapers; Obtain the structural control information of the icon matrix; The text description corresponding to the user instruction is used as the first generation condition, the style feature vector is used as the second generation condition, and the structural control information is used as the third generation condition. These are all input into the diffusion model to generate a stylized icon matrix that is consistent with the target style cockpit wallpaper style and maintains the original structure. The stylized icon matrix is ​​divided into multiple independent stylized application icons; The target style cockpit wallpaper and multiple stylized application icons were distributed and applied to the cockpit display interface.

2. The method for generating personalized cockpit themes according to claim 1, characterized in that, The implementation of generating target-style cockpit wallpapers corresponding to user commands includes: Receive user commands in voice format and convert them into text commands; Perform security audits on text commands; After approval, the text instructions are enhanced using a large language model to obtain enhanced prompts. The enhanced prompt words are input into the text-based image model to generate cockpit wallpapers in the target style. Perform image content security review on the target style cockpit wallpaper; After approval, the target style cockpit wallpaper will be stored on a cloud server.

3. The method for generating personalized cockpit themes according to claim 1, characterized in that, Multiple original application icons are concatenated into an icon matrix, including: Multiple original application icons with preset resolutions are arranged and stitched together according to an N*M grid to generate a combined matrix image, and the position coordinate information of each original application icon in the matrix image is recorded.

4. The method for generating personalized cockpit themes according to claim 1, characterized in that, The implementation of extracting style feature vectors from target style cockpit wallpapers includes: The target style cockpit wallpaper is encoded using the IP-Adapter image feature adapter to obtain a style feature vector, which is then used to inject style information into the denoising process of the diffusion model.

5. The method for generating personalized cockpit themes according to claim 1, characterized in that, The implementation of obtaining the structural control information of the icon matrix includes: Edge detection is performed on the icon matrix to obtain the outline structure map of all original application icons in the icon matrix; the structural control information is the outline structure map.

6. The method for generating personalized cockpit themes according to claim 1, characterized in that, The implementation of using structural control information as a third generation condition input into the diffusion model includes: The ControlNet control network is used to guide the generation process of the diffusion model by using the contour structure map as a control condition, so that the contour structure of each icon in the generated stylized icon matrix is ​​consistent with the corresponding original application icon.

7. The method for generating personalized cockpit themes according to claim 1, characterized in that, The implementation of dividing a stylized icon matrix into multiple independent stylized application icons includes: dividing the stylized icon matrix into independent icons corresponding to the original number of application icons based on position coordinate information, and performing image content security audits on each of the divided stylized application icons.

8. A device for generating personalized cockpit themes, applied to a vehicle-cloud collaborative system, employing the method for generating personalized cockpit themes as described in any one of claims 1 to 7, characterized in that, The device includes: The wallpaper generation module is used to generate cockpit wallpapers in the target style corresponding to user commands. The matrix building module is used to obtain multiple original application icons within the cockpit system and concatenate these icons into an icon matrix. The feature extraction module is used to extract style feature vectors based on the target style cockpit wallpaper; The structure acquisition module is used to acquire the structural control information of the icon matrix; The collaborative generation module is used to input the text description corresponding to the user's instruction as the first generation condition, the style feature vector as the second generation condition, and the structural control information as the third generation condition into the diffusion model to generate a stylized icon matrix that is consistent with the target style cockpit wallpaper style and maintains the original structure. The icon segmentation module is used to segment the stylized icon matrix into multiple independent stylized application icons; The theme application module is used to distribute and apply target-style cockpit wallpapers and multiple stylized application icons to the cockpit display interface.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The processor executes the program to implement the steps of the method for generating a cockpit personalization theme as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When a computer program is executed by a processor, it implements the steps of the method for generating a cockpit personalization theme as described in any one of claims 1 to 7.

Images