Three-dimensional animation production method based on AI deduction real-time rendering output

Abstract

The invention relates to the technical field of three-dimensional animation production, and discloses a three-dimensional animation production method based on AI deduction real-time rendering output, which comprises the following steps: S1, motion capture: in a motion capture stage, in combination with an AI technology, carrying out real-time analysis and optimization on original motion capture data by using a deep learning algorithm, and through a human body posture estimation technology, carrying out real-time analysis and optimization on the original motion capture data; the motion of the human body can be more accurately captured, and facial expressions and emotion changes are captured in real time in combination with an emotion recognition algorithm, so that more real and vivid animation role performance is provided; by optimizing and simplifying the three-dimensional animation production method, the production difficulty is remarkably reduced, the time is shortened, the cost is reduced, and children 'animations generally require high-frequency animation updating and high-quality visual effect to attract and maintain the interest of small audiences. The AI technology is integrated for action capture and emotion recognition, so that expressions and actions of roles can be captured and optimized in real time, and the roles can be expressed more naturally and vividly.

Description

Technical Field

[0001] The present invention relates to the technical field of three-dimensional animation production, and in particular to a three-dimensional animation production method based on AI deduction and real-time rendering output. Background Art

[0002] With the development of science and technology, 3D animation production is increasingly widely used in film, television, games and other fields. Traditional 3D animation production methods usually involve tedious key frame production, model building, texture mapping and other links, with long production cycles, high costs, and certain challenges in pursuing authenticity and naturalness.

[0003] Although existing motion capture technology can obtain real human movement and facial expression data, there are still some problems in its application in animation production. After traditional motion capture, producers need to spend a lot of time manually adjusting the animation to adapt to different scenes and effect requirements. In addition, real-time rendering and preview technology is also constantly developing, but how to effectively combine motion capture data with real-time rendering engines to improve production efficiency and output quality is still a problem to be solved. Based on this, technicians in this field propose a 3D animation production method based on AI deductive real-time rendering output to solve the above problems. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a 3D animation production method based on AI deduction and real-time rendering output, which solves the problems of long production cycle, high cost and poor authenticity in the 3D animation production process in the prior art.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a 3D animation production method based on AI deduction and real-time rendering output, comprising the following steps:

[0006] S1. Motion capture: In the motion capture stage, AI technology is combined with deep learning algorithms to analyze and optimize the original motion capture data in real time. Human posture estimation technology can more accurately capture human movements, and emotion recognition algorithms can be combined to capture facial expressions and emotional changes in real time, thereby providing more realistic and vivid animation character performances.

[0007] The S1 step specifically includes the following steps:

[0008] S101. Use personnel equipped with motion capture sensors to perform motion capture. These sensors may be inertial sensors, cameras, or other sensors to record the movements of athletes or actors. At the same time, the cameras can capture facial expressions and emotional changes;

[0009] The S1 step specifically includes the following steps:

[0010] S101. Ensure that the motion capture equipment (such as motion capture sensors, cameras, etc.) is in normal working condition so as to capture the movements and facial expressions of the personnel;

[0011] S102. Properly install the motion capture sensor on the person to ensure that the data of human body movement and facial expression can be accurately captured;

[0012] S104, calibrating and debugging the sensor to ensure that the sensor can accurately capture the movement trajectory and facial expression changes of the person, and at the same time requiring the person to perform predetermined movements and expressions, and starting the motion capture device to obtain original motion capture data;

[0013] S2. Use 3D scanners or cameras to collect data such as objects, light, materials, etc. in the scene, and process and optimize them. Use 3D modeling software to create and edit objects, environments, and characters in the scene, and then lay them out and organize them.

[0014] S3, 3D real-time rendering engine rendering: sending the adjusted motion capture animation to the 3D real-time rendering engine for real-time rendering and preview;

[0015] S4. Set appropriate materials and maps for objects and characters in the scene, including properties such as color, texture, and reflection. Use the skeletal system to create animations for characters and objects, then perform skeletal binding and weight adjustment to ensure that they can move and deform naturally. Design appropriate lighting effects according to scene requirements and set rendering parameters to achieve the desired visual effect.

[0016] S5. Animation generation: In the animation generation process, AI technology can be used to optimize animation effects and improve production efficiency. Generative adversarial networks (GANs) are used to generate more realistic lighting and material effects, or reinforcement learning algorithms are used to optimize the character's motion trajectory and action performance. At the same time, combined with deep learning technology, intelligent scene editing and automated animation generation can be achieved, thereby improving production efficiency and ensuring animation quality;

[0017] S6, storage and post-processing: storing the generated video animation and performing post-processing;

[0018] S7. Output and monitoring: During the rendering output stage, AI technology and graphics algorithms are used to optimize rendering speed and quality. Deep learning algorithms are used to perform real-time anti-aliasing operations to improve image quality and rendering speed.

[0019] Preferably, the step S1 specifically includes the following steps:

[0020] S101. Ensure that the motion capture equipment (such as motion capture sensors, cameras, etc.) is in normal working condition so as to capture the movements and facial expressions of the personnel;

[0021] S102. Properly install the motion capture sensor on the person to ensure that the data of human body movement and facial expression can be accurately captured;

[0022] S104, calibrating and debugging the sensor to ensure that the sensor can accurately capture the movement trajectory and facial expression changes of the person, and at the same time requiring the person to perform predetermined movements and expressions, and starting the motion capture device to obtain original motion capture data.

[0023] Preferably, the step S3 specifically includes the following steps:

[0024] S301, importing the adjusted and processed motion capture data into the animation editing environment of the three-dimensional real-time calculation engine;

[0025] S302, setting an animation scene in a three-dimensional real-time rendering engine, including background, lighting, camera position, etc., and importing a corresponding character model, applying the adjusted motion capture data to the corresponding character model, and making the character model perform motion performance according to the captured motion data;

[0026] S303, starting the real-time rendering function of the 3D real-time rendering engine, performing real-time rendering and previewing on the character model to which the motion capture data is applied, so as to observe the animation effect and make adjustments;

[0027] S304, adjusting the posture, expression, lighting and other parameters of the character model according to the real-time preview effect to achieve the desired visual effect;

[0028] S305. Production staff can interact with the character model during real-time rendering and preview, and provide feedback based on actual effects, thereby further optimizing the animation effects.

[0029] Preferably, the step S4 specifically includes the following steps:

[0030] S401, determine the background, scenery, props and other elements of the animation scene to ensure that they are consistent with the motion capture animation content and prepare for subsequent rendering;

[0031] S402, modeling the character in the motion capture data, including details such as muscles, clothing, facial expressions, etc., to make the character more vivid and realistic in the animation;

[0032] S403, adjusting the lighting effect according to the scene requirements and the context of the motion capture animation to improve the visual quality and atmosphere of the animation;

[0033] S404, adding special effects, such as smoke, fire, explosion, etc., to enhance the visual impact and realism of the motion capture animation, and performing rendering settings and optimization to ensure that the motion capture animation can present high-quality images and details during the rendering process;

[0034] S405. According to the content and plot of the motion capture animation, corresponding sound effects are added, such as ambient sound, background music, character dialogue, etc., to enrich the auditory experience of the animation, and post-processing is performed, including color correction, lens switching, special effects superposition, etc., to enhance the overall effect and viewing experience of the motion capture animation.

[0035] Preferably, the step S5 specifically includes the following steps:

[0036] S501, determining the background, scenery, props and other elements of the animation scene, ensuring that they are consistent with the motion capture animation content, and preparing for subsequent rendering, modeling the characters in the motion capture data, including details such as muscles, clothing, facial expressions, etc., to make the characters more vivid and realistic in the animation;

[0037] S502, adjusting the lighting effects according to the scene requirements and the context of the motion capture animation to improve the visual quality and atmosphere of the animation, and adding special effects such as smoke, fire, explosion, etc. to enhance the visual impact and realism of the motion capture animation;

[0038] S503, performing rendering settings and optimization to ensure that the motion capture animation can present high-quality images and details during the rendering process, and adding corresponding sound effects such as ambient sound, soundtrack, character dialogue, etc. according to the content and plot of the motion capture animation to enrich the auditory experience of the animation;

[0039] S504, finally post-processing is performed, including color correction, lens switching, special effects superposition, etc., to enhance the overall effect and viewing experience of the motion capture animation, and finally the processed motion capture animation is output to a video format and stored for subsequent playback, transmission and release.

[0040] Preferably, the post-processing in step S6 includes post-production and editing of audio and video.

[0041] Preferably, optimization and feedback are also included for optimizing and improving the video animation according to the monitoring results and user feedback to improve the quality and user satisfaction.

[0042] The present invention provides a 3D animation production method based on AI deduction and real-time rendering output. It has the following beneficial effects:

[0043] 1. The present invention significantly reduces the difficulty of production, shortens the time and reduces the cost by optimizing and simplifying the three-dimensional animation production method. Children's cartoons usually require high-frequency animation updates and high-quality visual effects to attract and maintain the interest of young audiences. By integrating AI technology for motion capture and emotion recognition, the expressions and movements of characters can be captured and optimized in real time, making the character performance more natural and vivid. The real-time rendering function allows the production team to quickly view the animation effects and make necessary adjustments, thereby improving production efficiency. This efficient production process enables the production team to launch more content in a shorter period of time, thereby meeting the needs of children's audiences for fresh and interesting content, while reducing the high costs caused by multiple modifications and re-rendering. This method enables even small production companies with limited budgets to create high-quality children's cartoons.

[0044] 2. The production method of the present invention integrates steps such as motion capture, real-time calculation engine, graphics processing and special effects setting. In the production of children's cartoons, the integration of efficient AI technology and real-time calculation engine greatly improves the quality of animation and audience experience. Children's cartoons require creative and highly delicate visual effects to attract and educate young audiences. The application of AI technology, including generative adversarial networks (GAN) and deep learning technology, can generate more realistic lighting, materials and scene effects to ensure that the visual effects of cartoons are more vivid and fascinating. In addition, through intelligent scene editing and automated animation generation, the production team can quickly achieve complex animation effects and scene settings, thereby completing higher-quality animation content in a short time. The application of this technology not only optimizes the animation production process, but also provides children audiences with a more interesting and rich visual experience, enhancing their sense of immersion and participation. Ultimately, this advanced production method improves the overall viewing and educational value of cartoons, so that it can better serve children's entertainment and education needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the decomposition process of the present invention;

[0046] Figure 2 It is a schematic diagram of the overall process of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] Example:

[0049] The embodiment of the present invention provides a 3D animation production method based on AI deduction and real-time rendering output, comprising the following steps:

[0050] S1. Motion capture: Use personnel equipped with motion capture sensors to perform motion capture to obtain raw motion capture data, which includes information such as human body movements and facial expressions for subsequent animation production;

[0051] Human pose estimation usually relies on deep learning models, especially convolutional neural networks (CNNs), which are able to extract complex features from raw image data. This process is divided into two main steps:

[0052] 1. Key point detection: The model predicts the positions of key points of the human body (such as shoulders, elbows, wrists, knees, etc.) by learning the spatial features in the image;

[0053] 2. Posture reconstruction: Based on the detected key points, the model further infers the overall posture of the human body, sometimes including the position and orientation in three-dimensional space.

[0054] Specifically, by using people equipped with motion capture sensors to perform motion capture, the movement and expression data of real human bodies can be accurately captured. This helps to improve the authenticity and naturalness of animation, making character movements more realistic and facial expressions more emotional.

[0055] The S1 step specifically includes the following steps:

[0056] S101. Ensure that the motion capture equipment (such as motion capture sensors, cameras, etc.) is in normal working condition so as to capture the movements and facial expressions of the personnel;

[0057] S102. Properly install the motion capture sensor on the person to ensure that the data of human body movement and facial expression can be accurately captured;

[0058] S104, calibrating and debugging the sensor to ensure that the sensor can accurately capture the movement trajectory and facial expression changes of the person, and at the same time requiring the person to perform predetermined movements and expressions, and starting the motion capture device to obtain original motion capture data.

[0059] S2, motion capture data processing: process and analyze the original motion capture data, generate motion capture animation, and adjust facial expressions;

[0060] Specifically, processing raw motion capture data can help remove possible noise and unwanted motion, resulting in more accurate and realistic animations. This helps improve the quality and realism of animations, making character movements more natural.

[0061] S3, 3D real-time rendering engine rendering: sending the adjusted motion capture animation to the 3D real-time rendering engine for real-time rendering and preview;

[0062] Specifically, through real-time rendering and preview, motion designers or production staff can immediately see the effects of the adjusted animation. This provides them with the opportunity for real-time feedback, allowing them to quickly adjust and optimize the motion to ensure that the desired effect is achieved. This greatly speeds up the production process and improves work efficiency.

[0063] The S3 step specifically includes the following steps:

[0064] S301, importing the adjusted and processed motion capture data into the animation editing environment of the three-dimensional real-time calculation engine;

[0065] S302, setting an animation scene in a three-dimensional real-time rendering engine, including background, lighting, camera position, etc., and importing a corresponding character model, applying the adjusted motion capture data to the corresponding character model, and making the character model perform motion performance according to the captured motion data;

[0066] S303, starting the real-time rendering function of the 3D real-time rendering engine, performing real-time rendering and previewing on the character model to which the motion capture data is applied, so as to observe the animation effect and make adjustments;

[0067] S304, adjusting the posture, expression, lighting and other parameters of the character model according to the real-time preview effect to achieve the desired visual effect;

[0068] S305. Production staff can interact with the character model during real-time rendering and preview, and provide feedback based on actual effects, thereby further optimizing the animation effects.

[0069] S4. Set appropriate materials and maps for objects and characters in the scene, including properties such as color, texture, and reflection. Use the skeletal system to create animations for characters and objects, then perform skeletal binding and weight adjustment to ensure that they can move and deform naturally. Design appropriate lighting effects according to scene requirements and set rendering parameters to achieve the desired visual effect.

[0070] The S4 step specifically includes the following steps:

[0071] S401, determine the background, scenery, props and other elements of the animation scene to ensure that they are consistent with the motion capture animation content and prepare for subsequent rendering;

[0072] S402, modeling the character in the motion capture data, including details such as muscles, clothing, facial expressions, etc., to make the character more vivid and realistic in the animation;

[0073] S403, adjusting the lighting effect according to the scene requirements and the context of the motion capture animation to improve the visual quality and atmosphere of the animation;

[0074] S404, adding special effects, such as smoke, fire, explosion, etc., to enhance the visual impact and realism of the motion capture animation, and performing rendering settings and optimization to ensure that the motion capture animation can present high-quality images and details during the rendering process;

[0075] S405. According to the content and plot of the motion capture animation, corresponding sound effects are added, such as ambient sound, background music, character dialogue, etc., to enrich the auditory experience of the animation, and post-processing is performed, including color correction, lens switching, special effects superposition, etc., to enhance the overall effect and viewing experience of the motion capture animation.

[0076] S5. Animation generation: Use AI technology to optimize and improve scenes, including using machine learning algorithms to train more accurate object recognition models, optimizing lighting and material generation algorithms, and using real-time rendering engines to render scenes in real time and output them as videos or real-time interactive applications; this includes object recognition model optimization, which includes:

[0077] Data collection and preprocessing

[0078] Collect a large number of diverse and representative scene images, covering different lighting, angles, occlusions and other factors.

[0079] The collected data is labeled, including object category, location, size and other information.

[0080] Preprocess the data, such as scaling, cropping, flipping, etc., to increase the generalization ability of the data

[0081] Algorithm selection and training

[0082] Use deep learning algorithms such as convolutional neural networks (CNN) for object recognition.

[0083] Use transfer learning techniques, such as feature extraction based on pre-trained ResNet, VGG and other models.

[0084] Model evaluation and optimization

[0085] Model performance was evaluated using metrics such as cross validation, precision, and recall.

[0086] To address the existing problems in the model, we perform operations such as parameter tuning and data enhancement to further improve recognition accuracy;

[0087] The S5 step specifically includes the following steps:

[0088] S501. Collect a large number of image data sets containing various scenes and objects, and mark and annotate the image data to indicate the location and category of each object, and use machine learning algorithms (such as convolutional neural networks) to train object recognition models to accurately identify and locate objects in the scene.

[0089] S502. Use technologies such as Generative Adversarial Networks (GAN) to generate more detailed and realistic materials. The application of GAN in 3D animation production mainly includes:

[0090] Lighting and material effect generation: Generate realistic lighting effects and material textures;

[0091] Specifically, physically based rendering (PBR): uses physically based rendering technology to simulate how light interacts with different materials in the real world, including reflection, refraction, scattering and absorption.

[0092] Global Illumination: Through ray tracing technology, it simulates the propagation path of light and produces natural lighting effects such as soft shadows, reflected and refracted light, ambient occlusion (AO), etc.

[0093] Dynamic Lighting: Calculate lighting changes in real time to adapt to dynamic scenes, such as moving light sources or changes in ambient lighting.

[0094] Lightmaps and Baking: Pre-calculate lighting information and store it in maps to optimize rendering performance, suitable for static scenes.

[0095] Material texture generation:

[0096] Texture synthesis: Use algorithms to automatically generate complex and natural textures such as wood, stone, skin, etc.

[0097] Procedural textures: Create infinitely detailed textures through mathematical functions and noise algorithms, which can be reused at different scales.

[0098] Example-based texture synthesis: learning and generating new textures from given texture samples, keeping the style and features of the original textures;

[0099] Motion data synthesis: Generate new motion data to enrich the motion library of animated characters. Specifically, generate new motion data:

[0100] Motion Blending: Combine different motion clips to create new motion sequences through interpolation and blending techniques.

[0101] Motion retargeting: Adapting the motion data of one character to another character with a different body structure.

[0102] Deep learning generation: Generate new action data using deep learning models such as generative adversarial networks (GANs) or variational autoencoders (VAEs).

[0103] Physics simulation: Simulate the character's movements based on the laws of physics to ensure the naturalness and realism of the movements.

[0104] Enrich the action library of animated characters:

[0105] Motion capture: Use motion capture technology to record the actor's movements and convert them into animation data.

[0106] Motion Editing: Providing tools that allow animators to adjust and optimize captured motion to suit different scenarios and needs;

[0107] Image enhancement: Image super-resolution and denoising are performed in the post-processing stage. Specifically,

[0108] Image Super Resolution:

[0109] Deep learning super-resolution: Use convolutional neural networks (CNNs) or other deep learning architectures to recover high-resolution details from low-resolution images.

[0110] Multi-frame fusion: Combine the information from multiple low-resolution images and synthesize a high-resolution image through an algorithm.

[0111] Frequency domain method: Process the image in the frequency domain and improve the resolution of the image by interpolating high-frequency components.

[0112] Denoising operation:

[0113] Deep learning denoising: Using deep learning models, especially convolutional neural networks, to learn to recover clean images from noisy images.

[0114] Non-local means denoising: Considers repetitive patterns in the image and removes noise by weighted averaging of non-local pixel values.

[0115] Wavelet transform denoising: The image is decomposed into coefficients of different scales through wavelet transform, and then the noise coefficients are thresholded.

[0116] S503, inputting the optimized scene data into the real-time rendering engine; setting parameters such as illumination, shadow, reflection, and camera viewing angle for the selected rendering engine;

[0117] S504: Perform real-time rendering and output it into a video file or embed it into a real-time interactive application as needed.

[0118] S6, storage and post-processing: storing the generated video animation and performing post-processing;

[0119] Specifically, the storage phase is the process of saving the generated video animation to a persistent storage device. This helps with long-term data preservation and backup to prevent accidental data loss. During the production cycle, it is possible to revisit the animation data for modification or re-editing, so an effective storage mechanism is essential for the long-term management of the project.

[0120] Post-processing in step S6 includes post-production and editing of audio and video.

[0121] S7. Output and monitoring: Use the video output device to output the video animation, and monitor the output video animation to ensure that the final effect meets the requirements.

[0122] The video output device in step S7 includes a display, a projector, a video converter, etc.

[0123] It also includes optimization and feedback, which is used to optimize and improve video animations based on monitoring results and user feedback to improve quality and user satisfaction.

[0124] More specifically, the present invention uses a ray tracing algorithm to simulate the propagation of light in a three-dimensional scene, and its formula is [R(t)=O+tD]

[0125] Among them, (O) is the starting point of the light, (D) is the direction vector of the light, and (t) is the parameter of the light, which can be any real number. When (t>0), it means that the direction of the light is from (O) along the positive direction of (D); when (t=0), it means that the starting point of the light is (O).

[0126] In summary, by optimizing and simplifying the 3D animation production method, the difficulty of 3D animation production is greatly reduced, the production time of 3D animation is shortened, and the development cost of 3D animation is reduced. By using the 3D real-time rendering engine for real-time rendering and preview, the producer can immediately see the adjusted animation effect, which helps to improve production efficiency and enables the producer to adjust and optimize the animation more flexibly. By integrating motion capture, real-time rendering engine, graphics processing and special effects setting, an efficient, realistic and high-quality animation production process is achieved, providing creators with more powerful tools and more flexible production processes.

[0127] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A 3D animation production method based on AI deduction and real-time rendering output, characterized in that: The following steps are involved: S1. Motion capture: In the motion capture stage, AI technology is combined with deep learning algorithms to analyze and optimize the original motion capture data in real time. Human posture estimation technology can more accurately capture human movements, and emotion recognition algorithms can be combined to capture facial expressions and emotional changes in real time, thereby providing more realistic and vivid animation character performances. The S1 step specifically includes the following steps: S101. Use personnel equipped with motion capture sensors to perform motion capture. These sensors may be inertial sensors, cameras, or other sensors to record the movements of athletes or actors. At the same time, the cameras can capture facial expressions and emotional changes; S102, after obtaining the original motion capture data, perform data preprocessing, including steps such as noise removal, data alignment and time synchronization, to ensure the accuracy and consistency of the data; S103, using deep learning algorithms to estimate human posture, which can analyze sensor data or video streams and identify key joints and posture information of the human body, such as body rotation, joint bending, etc.; S104, calibrating and debugging the sensor to ensure that the sensor can accurately capture the movement trajectory and facial expression changes of the person, and at the same time requiring the person to perform predetermined movements and expressions, and starting the motion capture device to obtain original motion capture data; S2. Use 3D scanners or cameras to collect data such as objects, light, materials, etc. in the scene, and process and optimize them. Use 3D modeling software to create and edit objects, environments, and characters in the scene, and then lay them out and organize them. S3, 3D real-time rendering engine rendering: sending the adjusted motion capture animation to the 3D real-time rendering engine for real-time rendering and preview; S4. Set appropriate materials and maps for objects and characters in the scene, including properties such as color, texture, and reflection. Use the skeletal system to create animations for characters and objects, then perform skeletal binding and weight adjustment to ensure that they can move and deform naturally. Design appropriate lighting effects according to scene requirements and set rendering parameters to achieve the desired visual effect. S5. Animation generation: In the animation generation process, AI technology can be used to optimize animation effects and improve production efficiency. Generative adversarial networks (GANs) are used to generate more realistic lighting and material effects, or reinforcement learning algorithms are used to optimize the character's motion trajectory and action performance. At the same time, combined with deep learning technology, intelligent scene editing and automated animation generation can be achieved, thereby improving production efficiency and ensuring animation quality; S6, storage and post-processing: storing the generated video animation and performing post-processing; S7. Output and monitoring: During the rendering output stage, AI technology and graphics algorithms are used to optimize rendering speed and quality. Deep learning algorithms are used to perform real-time anti-aliasing operations to improve image quality and rendering speed.

2. The method for producing 3D animation based on AI deduction and real-time rendering output according to claim 1, characterized in that: The S3 step specifically includes the following steps: S301, importing the adjusted and processed motion capture data into the animation editing environment of the three-dimensional real-time calculation engine; S302, setting an animation scene in a three-dimensional real-time rendering engine, including background, lighting, camera position, etc., and importing a corresponding character model, applying the adjusted motion capture data to the corresponding character model, and making the character model perform motion performance according to the captured motion data; S303, starting the real-time rendering function of the 3D real-time rendering engine, performing real-time rendering and previewing on the character model to which the motion capture data is applied, so as to observe the animation effect and make adjustments; S304, adjusting the posture, expression, lighting and other parameters of the character model according to the real-time preview effect to achieve the desired visual effect; S305. Production staff can interact with the character model during real-time rendering and preview, and provide feedback based on actual effects, thereby further optimizing the animation effects.

3. The method for producing 3D animation based on AI deduction and real-time rendering output according to claim 1, characterized in that: The S4 step specifically includes the following steps: S401, determine the background, scenery, props and other elements of the animation scene to ensure that they are consistent with the motion capture animation content and prepare for subsequent rendering; S402, modeling the character in the motion capture data, including details such as muscles, clothing, facial expressions, etc., to make the character more vivid and realistic in the animation; S403, adjusting the lighting effect according to the scene requirements and the context of the motion capture animation to improve the visual quality and atmosphere of the animation; S404, adding special effects, such as smoke, fire, explosion, etc., to enhance the visual impact and realism of the motion capture animation, and performing rendering settings and optimization to ensure that the motion capture animation can present high-quality images and details during the rendering process; S405. According to the content and plot of the motion capture animation, corresponding sound effects are added, such as ambient sound, background music, character dialogue, etc., to enrich the auditory experience of the animation, and post-processing is performed, including color correction, lens switching, special effects superposition, etc., to enhance the overall effect and viewing experience of the motion capture animation.

4. The method for producing 3D animation based on AI deduction and real-time rendering output according to claim 1, characterized in that: The step S5 specifically includes the following steps S501. Collect a large number of image data sets containing various scenes and objects, and mark and annotate the image data to indicate the location and category of each object, and use machine learning algorithms (such as convolutional neural networks) to train object recognition models to accurately identify and locate objects in the scene. S502. Use technologies such as generative adversarial networks (GAN) to generate more detailed and realistic materials. S503, inputting the optimized scene data into the real-time rendering engine; For the selected rendering engine, set parameters such as lighting, shadows, reflections, and camera perspectives; S504: Perform real-time rendering and output it into a video file or embed it into a real-time interactive application as needed.

5. The method for producing 3D animation based on AI deduction and real-time rendering output according to claim 1, characterized in that: The post-processing in step S6 includes post-production and editing of audio and video.

6. The method for producing 3D animation based on AI deduction and real-time rendering output according to claim 1, characterized in that: It also includes optimization and feedback, which is used to optimize and improve video animations based on monitoring results and user feedback to improve quality and user satisfaction.

Images