Image color changing method and system driven by multi-mode large model, and medium

Abstract

The invention provides a multi-mode large model driven image color changing method and system and a medium, and the method comprises the steps: obtaining to-be-processed vehicle image information, and carrying out the preprocessing of the vehicle image information, and the preprocessing comprises the adjustment of the size proportion, background color and image quality of the vehicle image information; based on the preprocessed vehicle image information, using a skeleton extraction model to obtain a vehicle skeleton image, the vehicle skeleton image being in a line draft form; calling a preset image generation model, inputting a vehicle skeleton image and auxiliary processing information, and generating a corresponding target vehicle image by the preset image generation model according to the vehicle skeleton image and the auxiliary processing information; the auxiliary processing information at least comprises rendering color information and image quality information. According to the method, multiple deep learning technologies are fully combined, and the image processing efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the automotive field, and in particular to an image color changing method, system and medium driven by a multi-modal large model. Background Art

[0002] Currently in the field of automotive design, the process by which designers convert car images into vehicle images of different colors includes the following steps: Using 3D modeling software, designers convert line drawings into three-dimensional models. This step involves creating the 3D geometry of the car and ensuring that every detail is accurately presented in the model. Then, materials and maps are added to the 3D model to give it a realistic appearance. Designers use texture mapping technology to simulate the surface materials of the car, such as metal paint, glass, and rubber. Next, the light source and camera angle are set to simulate the lighting effects in the real environment. Designers use rendering engines to generate high-quality renderings. This step requires adjusting lighting parameters, shadows, and reflection effects to ensure the realism of the final image. Then, image processing software is used to adjust and enhance the details of the rendering. Post-processing may include color correction, adding details, adjusting contrast and brightness, etc. Finally, the rendering is output as a high-resolution image or animation to show the design plan. These images can be used for product promotion, design review, or customer presentation. Through these steps, designers are able to transform the line drawings of the car into realistic renderings, thereby better demonstrating the design concept and product features.

[0003] However, the above method of converting car pictures into renderings of different colors has disadvantages such as high time cost, high labor cost and high complexity. This process requires a lot of time and professional skills, involves complex software and technology, and high-quality renderings require powerful computing power, which increases hardware costs. In addition, design modifications may require re-modeling and rendering, making it difficult to respond to changes quickly. Summary of the invention

[0004] In view of the defects in the prior art, the purpose of the present invention is to provide a multimodal large model driven image color changing method, system and medium, which combines multiple deep learning technologies, fully utilizes the learning and generation capabilities of the large model, and improves the image processing efficiency.

[0005] In order to achieve the above-mentioned object of the invention, the present invention adopts the following technical solution:

[0006] According to a first aspect of the present invention, there is provided a method for image color changing driven by a multimodal large model, comprising the following steps:

[0007] Acquire vehicle image information to be processed, and preprocess the vehicle image information, wherein the preprocessing includes adjusting the size ratio, background color, and image quality of the vehicle image information;

[0008] Based on the preprocessed vehicle image information, a vehicle skeleton image is obtained by using a skeleton extraction model, wherein the vehicle skeleton image is in the form of a line drawing;

[0009] Calling a preset image generation model, inputting a vehicle skeleton image and auxiliary processing information, and the preset image generation model generates a corresponding target vehicle image according to the vehicle skeleton image and the auxiliary processing information;

[0010] The auxiliary processing information at least includes rendering color information and image quality information.

[0011] Preferably, an auxiliary processing information library is provided to preset corresponding auxiliary processing information for different vehicle models.

[0012] According to a second aspect of the present invention, there is provided a multimodal large model driven image color changing system, comprising a preprocessing module, a skeleton extraction module, and a rendering module;

[0013] The preprocessing module is used to obtain the vehicle image information to be processed and preprocess the vehicle image information;

[0014] The skeleton extraction module is used to obtain a vehicle skeleton image based on the preprocessed vehicle image information using a skeleton extraction model, wherein the vehicle skeleton image is in the form of a line drawing;

[0015] The rendering module is used to call a preset image generation model, input a vehicle skeleton image and auxiliary processing information, and the preset image generation model generates a corresponding target vehicle image based on the vehicle skeleton image and the auxiliary processing information. The auxiliary processing information at least includes rendering color information and image quality information.

[0016] According to a third aspect of the present invention, there is provided a storage medium, which is a computer-readable storage medium, and which stores computer-executable instructions for causing a computer to execute the above-mentioned multimodal large model-driven image color changing method.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] The present invention combines multiple deep learning technologies and fully utilizes the learning and generation capabilities of large models to solve the problem that it takes a lot of time for designers to convert car pictures of different colors, thereby improving image processing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0020] Figure 1Schematic diagram of the process described in Example 1. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0022] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0023] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in the subsequent figures. In addition, all directional indications in this application (such as up, down, left, right, front, back, bottom...) are only used to explain the relative position relationship, movement, etc. between the components under a certain specific posture (as shown in the figures). If the specific posture changes, the directional indication will also change accordingly. Furthermore, the descriptions involving "first", "second", etc. in the application are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated.

[0024] Example 1

[0025] This embodiment provides a multi-modal large model driven image color changing method, based on the Stable Diffusion large model to change the color of the vehicle image, the vehicle image is converted into a vehicle skeleton and then converted into a vehicle image of different colors, model selection and positive and negative prompt word selection. Figure 1 As shown, the method provided in this embodiment includes the following steps:

[0026] In step S1, the vehicle image information to be processed is obtained and preprocessed. Specifically, in order to ensure the processing quality of the large model image, the size ratio, background color, and image quality of the vehicle image can be adjusted, for example, adjusted to a 16:9 ratio, and the background is adjusted to pure white, and the image is denoised.

[0027] In step S2, based on the pre-treated vehicle image information, a skeleton extraction model is utilized to obtain a vehicle skeleton image, wherein the vehicle skeleton image is in the form of a line drawing. Specifically, the skeleton extraction model used in the present embodiment is a self-training or existing image processing large model, and a vehicle picture with a 16:9 ratio and a pure white background is input, and converted into a line drawing picture with a black background and white lines. The vehicle skeleton image is extracted to retain the vehicle structural features during the image color change process, and the outline, edge and key features of the vehicle are represented in the form of a line drawing, so that the overall shape and detail structure of the vehicle can be effectively retained, and clear boundaries are provided for subsequent color rendering, so that the image generation model can more accurately fill the specified color to the corresponding position, avoid the situation of color overflow or incomplete filling, improve the accuracy of rendering, reduce the computational complexity, and improve the model generalization ability.

[0028] In step S3, the preset image generation model is called, and the vehicle skeleton image and auxiliary processing information are input. The preset image generation model generates a corresponding target vehicle image according to the vehicle skeleton image and the auxiliary processing information, and the auxiliary processing information at least includes rendering color information and image quality information.

[0029] Specifically, when the vehicle line drawing image is input into the preset image generation model for rendering, the preset image generation model can be a self-trained model or an existing image generation model, which is not limited in this embodiment. The input of the preset image generation model requires the input of auxiliary processing information as needed, and the auxiliary processing information includes positive prompt words, i.e. high image quality (such as best quality or masterpiece), required rendering color (such as green car), etc., and can also include reverse prompt words, i.e. low image quality (such as worst quality, low quality, normal quality, lowres, normal quality). The selection of positive prompt words and reverse prompt words should be adjusted according to different vehicle models, colors, lighting conditions and other factors to ensure the diversity and quality of the generated results.

[0030] Furthermore, in addition to the rendering color information and image quality information, more detailed descriptions can be introduced, such as vehicle body material information (such as metallic luster, matte) or ambient lighting conditions, to enhance the authenticity of the generated results and make the generated images more closely match the actual effects. Preferably, an auxiliary processing information library is set up to preset corresponding auxiliary processing information for different vehicle models, so as to facilitate batch processing of images of the same vehicle model.

[0031] This embodiment combines multiple deep learning technologies and fully utilizes the learning and generation capabilities of the large model to solve the problem that it takes a lot of time for designers to convert car pictures of different colors, thereby improving image processing efficiency.

[0032] Example 2

[0033] This embodiment provides a multi-modal large model driven image color changing system, which can realize a multi-modal large model driven image color changing method described in Example 1 through the cooperation of different modules. Specifically, this system includes a pre-processing module, a skeleton extraction module, and a rendering module. Among them:

[0034] The preprocessing module is used to obtain the vehicle image information to be processed and preprocess the vehicle image information;

[0035] The skeleton extraction module is used to obtain a vehicle skeleton image based on the preprocessed vehicle image information using a skeleton extraction model, wherein the vehicle skeleton image is in the form of a line drawing;

[0036] The rendering module is used to call a preset image generation model, input a vehicle skeleton image and auxiliary processing information, and the preset image generation model generates a corresponding target vehicle image based on the vehicle skeleton image and the auxiliary processing information. The auxiliary processing information at least includes rendering color information and image quality information.

[0037] It should be pointed out that the explanation of the implementation method and beneficial effects of the multimodal large model driven image color changing method in the above-mentioned embodiment 1 is also applicable to this embodiment, and will not be elaborated in detail here to avoid redundancy.

[0038] Example 3

[0039] This embodiment provides a storage medium, which is a computer-readable storage medium, and the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the image color changing method driven by a multimodal large model described in Example 1. The storage medium can be a volatile or non-volatile computer-readable storage medium.

[0040] It should be pointed out that the explanation of the implementation method and beneficial effects of the multimodal large model driven image color changing method in the above-mentioned embodiment 1 is also applicable to this embodiment, and will not be elaborated in detail here to avoid redundancy.

[0041] The above describes the specific embodiments of the present invention. Based on the above description, relevant personnel can make various changes and modifications without departing from the technical concept of the present invention.

Claims

1. A multi-modal large model driven image color changing method, characterized in that: The following steps are included: Acquire vehicle image information to be processed, and preprocess the vehicle image information, wherein the preprocessing includes adjusting the size ratio, background color, and image quality of the vehicle image information; Based on the preprocessed vehicle image information, a vehicle skeleton image is obtained by using a skeleton extraction model, wherein the vehicle skeleton image is in the form of a line drawing; Calling a preset image generation model, inputting a vehicle skeleton image and auxiliary processing information, and the preset image generation model generates a corresponding target vehicle image according to the vehicle skeleton image and the auxiliary processing information; The auxiliary processing information at least includes rendering color information and image quality information.

2. The image color changing method driven by a multi-modal large model according to claim 1, characterized in that: Set up an auxiliary processing information library and preset corresponding auxiliary processing information for different vehicle models.

3. A multi-modal large model driven image color changing system, characterized in that: Contains preprocessing module, skeleton extraction module, and rendering module; The preprocessing module is used to obtain the vehicle image information to be processed and preprocess the vehicle image information; The skeleton extraction module is used to obtain a vehicle skeleton image based on the preprocessed vehicle image information using a skeleton extraction model, wherein the vehicle skeleton image is in the form of a line drawing; The rendering module is used to call a preset image generation model, input a vehicle skeleton image and auxiliary processing information, and the preset image generation model generates a corresponding target vehicle image based on the vehicle skeleton image and the auxiliary processing information. The auxiliary processing information at least includes rendering color information and image quality information.

4. A storage medium, the storage medium being a computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the multi-modal large model-driven image color changing method as claimed in claim 1 or 2.

Images