Animation real-time capturing and repairing method and system

Abstract

The invention relates to the technical field of animation processing, in particular to an animation real-time capturing and repairing method and system, and the method comprises the following steps: calculating the pixel change rate between continuous animation frames based on the time sequence data of the animation, determining the change threshold values of key frames and non-key frames, dynamically adjusting image collection parameters according to the change threshold values, and carrying out the real-time capturing and repairing of the animation. And the frame difference before and after adjustment is compared and analyzed. According to the method, through analysis of the animation time sequence data and real-time calculation of the pixel change rate, the continuity between the animation frames is effectively monitored and real-time adjustment is performed, so that the accuracy and efficiency of animation production are improved, visual errors are corrected in real time in the production stage, the high-cost later-stage repair requirement is reduced, and the production efficiency is improved. The frame-by-frame real-time semantic segmentation and visual element identification further enhance the accurate control of the animation content, ensure the image quality of each frame, and ensure the consistency and visual effect of the animation by refining the image quality and synchronously repairing the frames.

Description

Technical Field

[0001] The present invention relates to the field of animation processing technology, and in particular to a method and system for real-time capturing and repairing animation. Background Art

[0002] Animation processing technology mainly focuses on animation production, optimization and interactive technology, including traditional 2D animation to modern 3D animation production technology, covering character modeling, motion capture, visual effects generation, and real-time image repair and enhancement. In this technical field, important technological advances include real-time rendering technology, animation simulation, and the interaction between characters and the environment. Animation processing technology makes animation production more efficient and dynamic, providing a richer and more realistic visual experience, especially in the fields of film, video games and virtual reality.

[0003] Among them, the real-time animation capture and repair method refers to the use of computer vision and image processing technology to identify and repair technical or visual errors in the animation in real time during the animation production process. It usually involves the use of dynamic capture technology to monitor the movement and performance of the animation, identify unnatural movements or distorted animation parts, and then make corrections immediately. The purpose is to correct errors instantly during the production stage, avoid costly post-repairs, and ensure the continuity and visual effects of the animation during creation and playback.

[0004] Existing technologies are unable to promptly detect and correct subtle unnatural movements or animation distortions during the animation production process, resulting in the production team having to invest a lot of time in later stages to make modifications, which not only increases costs but also prolongs the production cycle. In addition, the lack of an effective real-time adjustment mechanism results in improper handling of frame continuity and visual consistency, affecting the overall viewing experience of the animation. For example, failure to adjust frame parameters in a timely manner in dynamic scenes causes the audience to experience picture incoherence, reducing the immersion of the animation. This technical shortcoming is particularly prominent in fields that have strict requirements for high-quality animation output, affecting the quality of the animation. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a method and system for real-time capture and restoration of animation.

[0006] In order to achieve the above object, the present invention adopts the following technical scheme: a method for real-time capture and restoration of animation, comprising the following steps:

[0007] S1: Based on the animation time series data, the pixel change rate between consecutive animation frames is calculated, the change threshold between key frames and non-key frames is determined, the image acquisition parameters are dynamically adjusted according to the change threshold, and the rendering parameter optimization configuration is obtained by comparing and analyzing the frame differences before and after the adjustment;

[0008] S2: Based on the rendering parameter optimization configuration, the collected animation frames are semantically segmented in real time, key visual elements in the frames are identified frame by frame, including character outlines and backgrounds, and image quality is dynamically adjusted to obtain visual correction records;

[0009] S3: Based on the visual correction record, compare the quality indicators of the original frame and the repaired frame, including clarity and color accuracy, perform image quality evaluation, and adjust the image repair parameters according to the evaluation results to obtain frame quality refinement information;

[0010] S4: Synchronize the frame quality refinement information with the original animation sequence, check the continuity and visual consistency on the timeline through time code comparison, and adjust the frame interpolation process according to the timeline calibration result to optimize the smoothness and visual effect of the animation playback, and obtain the animation frame repair result.

[0011] The present invention is improved in that the step of determining the change threshold between the key frame and the non-key frame is specifically as follows:

[0012] S111: Based on the time series data of the animation, the pixel change rate between consecutive animation frames is calculated using the formula:

[0013]

[0014] Get the average rate of change V, where p i，t represents the color value of the i-th pixel at time t, p i，t +1 represents the color value of the i-th pixel at time t+1, and N represents the total number of pixels in the image;

[0015] S112: using percentiles, determining a threshold of the average change rate, and distinguishing between key frames and non-key frames; if the average change rate exceeds the threshold, the frame is regarded as a key frame, and a change threshold determination result is obtained.

[0016] The present invention is improved in that the step of obtaining the rendering parameter optimization configuration is specifically as follows:

[0017] S121: Based on the change threshold, adjusting image acquisition parameters, including frame interval and exposure settings, and capturing key frame changes to obtain adjusted acquisition parameter settings;

[0018] S122: Based on the adjusted acquisition parameter settings, reacquire the animation frame, and calculate the frame data difference before and after the adjustment, using the formula:

[0019]

[0020] Get the frame difference analysis result, where QC is the average pixel difference, p i，new Represents the color value of the i-th pixel after adjustment, pi，old represents the color value of the i-th pixel before adjustment, and N represents the total number of pixels in the image;

[0021] S123: Based on the frame difference analysis result, the compression ratio and resolution of the rendering parameters are continuously adjusted to match the optimal quality and efficiency balance of the animation rendering, and an optimized configuration of the rendering parameters is obtained.

[0022] The present invention is improved in that the step of identifying the key visual elements in the frame is specifically as follows:

[0023] S211: Based on the rendering parameter optimization configuration, perform animation frame acquisition, and verify whether each frame image meets a predetermined quality standard to obtain an optimized animation frame;

[0024] S212: Perform real-time semantic segmentation on the optimized animation frame using the formula:

[0025]

[0026] Get the segmentation data set, where SU i represents the segmentation result of the i-th pixel, xu ij is the value of the i-th pixel on the j-th feature, wu j is the weight of the jth feature, and M is the total number of features;

[0027] S213: Based on the segmented data set, extract and identify key visual elements in the frame, including the outline of the person and the background, and separate the outline of the person and the background to obtain a visual element recognition record.

[0028] The present invention is improved in that the steps of obtaining the visual correction record are specifically as follows:

[0029] S221: Based on the key visual elements, preliminary visual correction is performed on the image, including adjusting light and dark contrast, color saturation and sharpening, to obtain corrected image data;

[0030] S222: Performing a quality assessment on the corrected image data, assessing the quality difference between the correction effect and the original image, and obtaining a correction effect score;

[0031] S223: Based on the correction effect score, continuously adjust the image correction parameters, and optimize the image processing process to obtain a visual correction record.

[0032] The present invention is improved in that the step of comparing the quality indexes of the original frame and the repaired frame is specifically as follows:

[0033] S311: extracting data of the original frame and the repaired frame based on the visual correction record, including pixel brightness, color distribution and contrast attributes, to obtain original and repaired frame data sets;

[0034] S312: Analyze the original and restored frame data sets, calculate the clarity and color accuracy through the peak signal-to-noise ratio and the color fidelity index, and obtain preliminary comparative data of the image quality;

[0035] S313: Based on the preliminary comparison data of the image quality, the formula is used:

[0036]

[0037] Evaluate the quality of the image and obtain the image quality index QI, where WO 1 and WO 2 It is a weight parameter used to adjust the criticality of PSNR and CFI in quality assessment. PSNR is the peak signal-to-noise ratio and CFI is the color fidelity.

[0038] The present invention is improved in that the step of obtaining the frame quality refinement information is specifically as follows:

[0039] S321: According to the evaluation result, identifying the restoration parameters that cause the image quality degradation problem, the problem includes insufficient clarity or color distortion, and obtaining a degradation problem identification result;

[0040] S322: Based on the degradation problem identification result, adjust the image restoration parameters, including sharpening degree, color correction setting and contrast adjustment, evaluate the image quality again, and verify the adjustment effect to obtain frame quality refinement information.

[0041] The present invention is improved in that the steps of obtaining the animation frame repair result are specifically as follows:

[0042] S411: Synchronizing the frame quality refinement information with the original animation sequence, verifying the correspondence between each repaired frame and the original frame on the time axis through time code comparison, and obtaining calibrated time code data;

[0043] S412: Analyze the calibrated time code data to check the continuity and visual consistency on the time axis, using the formula:

[0044]

[0045] Get the visual consistency evaluation result, where VCI is the visual consistency index, VF i Represents the visual quality index of the i-th frame, VF i-1 Represents the visual quality index of the i-1th frame, N F Represents the total number of frames, abs(VF i -VF i-1 ) is the absolute value of the visual quality difference between the i-th frame and its previous frame i-1;

[0046] S413: Based on the visual consistency evaluation result, the frame interpolation process is adjusted to optimize the smoothness and visual effect of the animation playback, including adjusting the frame rate, inserting or deleting the target frame, and obtaining the animation frame repair result.

[0047] A real-time animation capture and restoration system, the system comprising:

[0048] The frame change rate analysis module calculates the pixel change rate between consecutive animation frames based on the animation time series data, determines the change threshold between key frames and non-key frames, dynamically adjusts the image acquisition parameters, and obtains the optimal configuration of rendering parameters by comparing and analyzing the frame differences before and after the adjustment;

[0049] The visual element recognition module performs real-time semantic segmentation on the collected animation frames based on the optimized configuration of the rendering parameters, recognizes key visual elements in the frames frame by frame, and dynamically adjusts the image quality to obtain a visual correction record;

[0050] The image quality tuning module compares the quality indicators of the original frame and the repaired frame based on the visual correction record, including clarity and color accuracy, evaluates the image quality, and adjusts the image repair parameters according to the evaluation results to obtain frame quality refinement information;

[0051] The timeline alignment module synchronizes the original animation sequence based on the frame quality refinement information, checks the continuity and visual consistency on the timeline through time code comparison, and obtains a timeline calibration result;

[0052] The playback optimization module adjusts the frame interpolation process based on the time axis calibration result, optimizes the smoothness and visual effect of the animation playback, and obtains the animation frame repair result.

[0053] Compared with the prior art, the advantages and positive effects of the present invention are:

[0054] In the present invention, through the analysis of animation time series data and the real-time calculation of pixel change rate, the continuity between animation frames is effectively monitored and adjusted in real time, thereby improving the accuracy and efficiency of animation production, and instantly correcting visual errors in the production stage, reducing the need for costly post-production repairs. Real-time semantic segmentation and visual element recognition frame by frame further enhance the precise control of animation content and ensure the image quality of each frame. By refining the image quality and synchronously repairing the frames, the continuity and visual effects of the animation are ensured, and the playback smoothness is optimized. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A flowchart of a method for real-time capture and restoration of animation is proposed for the present invention;

[0056] Figure 2A flow chart for determining a change threshold value between key frames and non-key frames in the present invention;

[0057] Figure 3 A flowchart for obtaining the optimized configuration of rendering parameters in the present invention;

[0058] Figure 4 is a flow chart for identifying key visual elements in a frame in the present invention;

[0059] Figure 5 A flowchart for obtaining visual correction records in the present invention;

[0060] Figure 6 A flow chart showing comparison of quality indicators of original frames and repaired frames in the present invention;

[0061] Figure 7 This is a flow chart for obtaining frame quality refinement information in the present invention;

[0062] Figure 8 This is a flow chart for obtaining the animation frame restoration results in the present invention. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0064] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating positions or positional relationships, are based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0065] Example

[0066] See also Figure 1 The present invention provides a technical solution: a method for real-time capture and repair of animation, comprising the following steps:

[0067] S1: Based on the animation time series data, the pixel change rate between consecutive animation frames is calculated, the change threshold between key frames and non-key frames is determined, the image acquisition parameters are dynamically adjusted according to the change threshold, and the rendering parameter optimization configuration is obtained by comparing and analyzing the frame differences before and after the adjustment;

[0068] S2: Based on the optimized configuration of rendering parameters, the collected animation frames are semantically segmented in real time, key visual elements in the frames are identified frame by frame, including the outline of the characters and the background, and the image quality is dynamically adjusted to obtain visual correction records;

[0069] S3: Based on the visual correction record, the quality indicators of the original frame and the repaired frame, including clarity and color accuracy, are compared to perform image quality evaluation, and the image restoration parameters are adjusted according to the evaluation results to obtain frame quality refinement information;

[0070] S4: Synchronize the frame quality refinement information with the original animation sequence, check the continuity and visual consistency on the timeline through time code comparison, and adjust the frame interpolation process according to the timeline calibration result to optimize the smoothness and visual effect of the animation playback and obtain the animation frame repair result.

[0071] The rendering parameter optimization configuration includes sampling rate adjustment information, resolution standard, and dynamic range setting. The visual correction records are specifically visual element classification, boundary clarity, and background isolation effect. The frame quality refinement information includes texture recovery, color balance, and brightness optimization indicators. The animation frame repair results specifically refer to the playback sequence continuity, visual contrast calibration results, and frame interpolation efficiency.

[0072] See also Figure 2 , the specific steps for determining the change threshold between key frames and non-key frames are:

[0073] S111: Based on the time series data of the animation, the pixel change rate between consecutive animation frames is calculated using the formula:

[0074]

[0075] The average change rate V is obtained to quantify the degree of pixel change between consecutive frames, where p i，t represents the color value of the i-th pixel at time t, p i，t+1 represents the color value of the i-th pixel at time t+1, N represents the total number of pixels in the image, which is used to calculate the average value;

[0076] There is an animation sequence in which each frame has only 3 pixels, and the pixel value is p at time t i，t = [100, 150, 200] and p at time t+1 i，t+1 =[110, 140, 210], then the square difference is calculated as:

[0077] ∑(p i，t -p i，t+1 ) 2 =(100-110) 2 +(150-140) 2 +(200-210)2 =100+100+100=300;

[0078] Then, divide this value by the number of pixels N = 3 to get the average rate of change:

[0079]

[0080] The result shows that the average change rate of each pixel is 100, which means that between the two frames, each pixel has changed an average of 100 square differences. This degree of change is sufficient to regard the current frame as a key frame. Based on this value, a threshold can be set to identify all key frames.

[0081] S112: using percentiles, determining a threshold of the average change rate, and distinguishing between key frames and non-key frames; if the average change rate exceeds the threshold, the frame is regarded as a key frame, and a change threshold determination result is obtained;

[0082] Using the percentile calculation method, the average change rate data of continuous animation frames are sorted according to the numerical size, the average change rate value corresponding to each percentile is determined, and the change rate corresponding to the 95th percentile in the sorted data is extracted as the preliminary threshold benchmark. In the calculation process, all the average change rate values ​​are compared one by one to determine the position of each value in the overall sorting, and then the percentile value is accurately determined by the cumulative frequency method. Then, the effectiveness of the selected percentile threshold is further verified. Using multiple groups of test animation data, frame by frame comparison and recording of the frame number ratio within and outside the threshold range are used to ensure that the set percentile threshold is applicable to scene changes of different animations, and to obtain the change threshold determination result.

[0083] See also Figure 3 , the specific steps for obtaining the rendering parameter optimization configuration are:

[0084] S121: Based on the change threshold, adjust the image acquisition parameters, including frame interval and exposure settings, and capture key frame changes to obtain adjusted acquisition parameter settings;

[0085] First, extract the pixel data of the current frame and compare it with the change threshold of the previous and next frames. Dynamically adjust the value of the frame interval according to the amount of change to reduce redundant acquisition of low-change areas. Next, analyze the overall lighting conditions of the animation, and adjust the exposure settings by counting the brightness distribution of pixels in each frame to ensure that the light intensity and details of the change area are within the acquisition range. During the process, it is necessary to ensure that the adjusted frame interval and exposure value maintain continuity between frames, while avoiding the omission of key frame information. Finally, generate a set of optimized acquisition parameter configurations to ensure that the acquired animation frames contain both key information and reduce redundancy.

[0086] S122: Based on the adjusted acquisition parameter settings, reacquire the animation frame and calculate the frame data difference before and after the adjustment using the formula:

[0087]

[0088] The frame difference analysis results are obtained, where QC is the average pixel difference, which is used to quantify the degree of change of frame data after adjusting the image acquisition parameters, and p i，new Represents the color value of the i-th pixel after adjustment, p i，old represents the color value of the i-th pixel before adjustment, and N represents the total number of pixels in the image;

[0089] There are 2 pixels in the animation frame, and the pixel values ​​before and after adjustment are:

[0090] p 1，old =100, p 1，new =110;

[0091] and

[0092] p 2，old =150, p 2，new =140;

[0093] Substitute the formula into the following calculation:

[0094]

[0095] The results show that the average color value change of each pixel is 10, which reflects the actual impact of image acquisition parameter adjustment on pixel values, verifies the effectiveness of the adjustment, and shows that adjusting acquisition parameters can significantly affect pixel data, which helps to further optimize parameters according to actual conditions.

[0096] S123: Based on the frame difference analysis result, continuously adjust the compression ratio and resolution of the rendering parameters to match the optimal quality and efficiency balance of the animation rendering, and obtain the optimized configuration of the rendering parameters;

[0097] In-depth adjustments are made to key areas in the frame data. The maximum and average values ​​of changes in frame differences are calculated to dynamically adjust the compression ratio and resolution. The compression ratio is adjusted based on the maximum value of key pixel changes, and details are retained by reducing the compression ratio in high-change areas. The resolution is adjusted based on the overall average difference value, and file size is reduced by increasing the resolution compression ratio in low-change areas. Finally, the adjusted frames are rendered one by one, and the specific impact of compression ratio and resolution adjustment on rendering efficiency is analyzed. Based on the results, optimization is repeated until a balance is reached between the details of the changing areas and the overall compression efficiency, completing the rendering parameter optimization configuration.

[0098] See also Figure 4, the identification steps of key visual elements in the frame are as follows:

[0099] S211: Based on the optimized configuration of rendering parameters, perform animation frame acquisition, and verify whether each frame image meets a predetermined quality standard to obtain an optimized animation frame;

[0100] In the process of collecting animation frames, it is necessary to ensure that the image quality of each frame meets the established standards. First, the resolution and color accuracy of each frame are verified. The resolution can be verified by checking whether the number of pixels is consistent with the set resolution. At the same time, a standardized color space mapping method is used for color accuracy to quantify and compare the color value of each pixel. Then, the stability of the frame rate is tested multiple times during the collection process to ensure that the time interval between frames is stable and to avoid inconsistency in the collected data due to frame rate fluctuations. Finally, after verifying the parameters, only frames that meet the quality standards are retained and stored to obtain optimized animation frames. The frames can be directly used for subsequent image segmentation and content analysis.

[0101] S212: Real-time semantic segmentation of the optimized animation frames using the formula:

[0102]

[0103] Get the segmentation data set, where SU i Indicates the segmentation result of the i-th pixel, with a value range of 0 to 1, which is used to indicate the probability that the pixel belongs to the foreground (such as the outline of a person). ij is the value of the i-th pixel on the j-th feature. The feature may include the color value, texture information, position information, etc. of the pixel. j is the weight of the jth feature. The weight is learned through training data and determines the importance of each feature in the segmentation decision. M is the total number of features.

[0104] There is a pixel in a frame image, and the probability of belonging to the foreground is calculated. It is known that its value in the color value feature is 180, the texture information is 0.5, the position information is 0.2, and the feature weights are 0.8, 0.1, and 0.1 respectively. Substitute the values ​​into the formula:

[0105]

[0106] SU i ≈1;

[0107] This result indicates that the pixel almost certainly belongs to the foreground, demonstrating its efficiency and accuracy in identifying key visual elements.

[0108] S213: extracting and identifying key visual elements in the frame, including a person outline and a background, based on the segmented data set, and separating the person outline and the background to obtain a visual element recognition record;

[0109] The segmentation results of each frame in the data set are analyzed pixel by pixel. First, the pixels are classified according to the probability value of the segmentation result. The pixels with probability values ​​higher than a specific threshold are marked as human contours, and those below the threshold are marked as backgrounds. Then, the boundary detection is performed on the area of ​​the human contour, and the closure and integrity of the contour are calculated. Isolated pixels and irregular edges are further removed. Subsequently, the human contour is refined using the position distribution information. At the same time, texture recognition and color statistics are performed on the background area to ensure the accuracy of background classification. The above processing finally completes the separation of the human contour and the background, and obtains the visual element recognition record.

[0110] See also Figure 5 , the specific steps for obtaining visual correction records are:

[0111] S221: Based on key visual elements, preliminary visual correction is performed on the image, including adjusting light and dark contrast, color saturation and sharpening, to obtain corrected image data;

[0112] First, it is necessary to adjust the brightness of the identified key visual element areas. This can be done by calculating the difference between the average brightness of all pixels and the target brightness value, and evenly distributing the difference value to each pixel, thereby correcting the overall brightness of the image. Then, it is necessary to adjust the color saturation by extracting the color channel value of each pixel, calculating its saturation coefficient, and appropriately adjusting the color channel value to ensure that the color is more vivid and natural. Finally, image sharpening is performed, and edge enhancement technology is used to extract important boundary information in the image. At the same time, subtle noise is suppressed to highlight the clarity and detail of the image, completing the preliminary visual correction of the image and obtaining the corrected image data.

[0113] S222: Performing a quality assessment on the corrected image data, assessing the quality difference between the correction effect and the original image, and obtaining a correction effect score;

[0114] When evaluating the quality of corrected image data, it is necessary to select appropriate evaluation criteria, such as using peak signal-to-noise ratio and image structure similarity to compare images in multiple aspects. First, statistical analysis is performed on the difference between each pixel of the corrected image and the original image, and the overall quality difference between the corrected image and the original image is evaluated by calculating the mean and variance of the error. Next, the details of the corrected image and the original image are analyzed, including the correlation between brightness, color, and structure, to generate a quality score for each frame of the image. Finally, the scores of each frame are integrated into an overall score as evaluation data for the correction effect.

[0115] S223: Based on the correction effect score, continuously adjust the image correction parameters and optimize the image processing process to obtain a visual correction record;

[0116] When further adjusting the image correction parameters, it is necessary to combine the specific scoring results in the quality assessment. First, analyze whether the brightness correction parameters meet the scoring requirements. For the parts that do not meet the standards, you can expand the brightness adjustment range or refine the graded adjustment. Then analyze the effectiveness of the color saturation correction. For the parts with oversaturation or undersaturation, re-optimize the adjustment coefficient of the color channel and dynamically adjust the boundary range of the saturation application area. Finally, optimize the image sharpening parameters, especially reduce the sharpening intensity in areas with more details and improve the sharpening effect in areas with less boundary information. After a continuous adjustment and optimization process, the image correction parameters are finally fixed and a complete visual correction record is recorded.

[0117] See also Figure 6 ,The steps of comparing the quality indicators of the original frame and the repaired frame are as follows:

[0118] S311: extracting data of the original frame and the repaired frame based on the visual correction record, including pixel brightness, color distribution, and contrast attributes, to obtain original and repaired frame data sets;

[0119] By analyzing each frame of the image, key data are extracted, including pixel brightness, color distribution and contrast properties. First, the pixel brightness is statistically divided into regions, and the histogram equalization method is used to calculate the uniformity and concentration of the brightness in each region. Then the color distribution is analyzed. The breadth and uniformity of the color distribution are evaluated by splitting the RGB channels and calculating the color histogram of each channel. Finally, the contrast property is calculated. The gradient amplitude is calculated for the areas with large brightness differences in each frame of the image, and the contrast distribution is comprehensively statistically calculated. The original and repaired frame data sets containing pixel brightness, color distribution and contrast properties are obtained.

[0120] S312: Analyze the original and restored frame data sets, calculate the clarity and color accuracy through the peak signal-to-noise ratio and color fidelity index, and obtain preliminary comparative data of image quality;

[0121] First, the peak signal-to-noise ratio is calculated. The PSNR value is calculated using a fixed formula by obtaining the square mean of the pixel difference between the original frame and the repaired frame. Then, the color fidelity index is calculated. By extracting the color information of each pixel in the frame, the color deviation analysis in spatial and temporal dimensions is performed. Finally, the color difference between the repaired frame and the original frame is quantified by the distribution coincidence to obtain a specific CFI value. Combined with the multi-frame results of clarity and color accuracy, preliminary comparative data of image quality is generated.

[0122] S313: Based on preliminary comparative data of image quality, the formula is used:

[0123]

[0124] Evaluate the quality of the image and obtain the image quality index QI, where WO 1 and WO 2 It is a weight parameter used to adjust the relative importance of PSNR and CFI in quality assessment. PSNR is the peak signal-to-noise ratio, a commonly used image quality assessment indicator that mainly measures the clarity of the image. CFI is the color fidelity, which measures the color accuracy between the restored frame and the original frame. A high CFI value indicates a small color deviation and high color fidelity.

[0125] The PSNR value of the original frame is 30dB, the PSNR value of the repaired frame is increased to 35dB, the CFI value is increased from 0.85 to 0.9, and the weight parameter WO 1 and WO 2 All are 1, substitute into the formula for calculation:

[0126]

[0127] The results show that the restored frames have an image quality index of 17.95 after comprehensively considering clarity and color fidelity, indicating that the restoration process significantly improves the overall quality of the image.

[0128] See also Figure 7 , the specific steps for obtaining frame quality refinement information are:

[0129] S321: identifying, according to the evaluation result, restoration parameters that cause image quality degradation problems, the problems including insufficient clarity or color distortion, and obtaining degradation problem identification results;

[0130] Analyze the source of image quality degradation problems, focus on the pixel value distribution in the blurred area of ​​the image and the statistical range of color deviation by comparing the clarity index and color fidelity index. For the problem of insufficient clarity, extract the brightness gradient value and edge detection information in the blurred area, and compare it with the image standard to determine the specific problem area. For the color distortion problem, calculate the deviation value between the color value of the pixel and the target color gamut, and count the number of pixels that exceed the color deviation threshold. Through the above analysis, accurately identify the parameter setting problems that lead to image quality degradation, including specific insufficient sharpening, contrast imbalance and color correction parameter errors, and finally obtain the degradation problem identification result.

[0131] S322: adjusting image restoration parameters according to the degradation problem identification result, including sharpening degree, color correction setting and contrast adjustment, evaluating the image quality again, and verifying the adjustment effect to obtain frame quality refinement information;

[0132] The degree of sharpening is reconfigured, and the appropriate sharpening intensity parameters are set according to the edge pixel gradient value of the blurred area. At the same time, the color correction settings are adjusted, and the threshold of color deviation is used as a reference. The pixel color values ​​that exceed the threshold are interpolated or directly corrected. For contrast problems, the pixel brightness distribution is adjusted by the histogram equalization method to ensure that the grayscale dynamic range of the image is optimal. After the above adjustments are completed, the clarity and color fidelity indicators of the adjusted repaired frame are calculated again, and the adjustment effect is verified by comparing with the original frame data, and finally the frame quality refinement information is generated.

[0133] See also Figure 8 , the specific steps for obtaining the animation frame repair results are:

[0134] S411: Synchronize the frame quality refinement information with the original animation sequence, verify the correspondence between each repaired frame and the original frame on the time axis through time code comparison, and obtain calibrated time code data;

[0135] The time codes of the original frames and the repaired frames are extracted frame by frame and converted into a standardized timeline unit format. The differences between the original time codes and the repaired time codes are then compared, and the parameters of any frame markers with offsets are adjusted. The adjustment method corrects the time offset of the repaired frames according to the magnitude of the time deviation to ensure that the repaired frames strictly match the original frames on the timeline, and generates calibrated time code data to provide basic data for subsequent analysis.

[0136] S412: Analyze the calibrated time code data to check the continuity and visual consistency on the time axis, using the formula:

[0137]

[0138] The visual consistency evaluation results are obtained, where VCI is the visual consistency index, which is used to quantify the visual continuity and consistency between animation frames, and VF i Represents the visual quality index of the i-th frame, which can be any quantitative image quality metric, such as brightness, contrast or color accuracy, etc. VF i-1 Represents the visual quality index of the i-1th frame, N F Represents the total number of frames, abs(VF i -VF i-1 ) is the absolute value of the visual quality difference between the i-th frame and its previous frame i-1;

[0139] The animation frame has a total of 1000 pixels. Before and after the adjustment, a certain pixel value changes from 120 to 130, and the rest do not change much, that is:

[0140]

[0141] The result shows that the average pixel difference is very small, indicating that the effect of the acquisition parameter adjustment is as expected and maintains the overall stability and quality of the animation.

[0142] S413: Based on the visual consistency evaluation result, the frame interpolation process is adjusted to optimize the smoothness and visual effect of the animation playback, including adjusting the frame rate, inserting or deleting the target frame, and obtaining the animation frame repair result;

[0143] By analyzing the visual consistency results of each frame, the frame rate optimization and interpolation strategies are determined. First, a cluster analysis is performed on the visual consistency of the frames, and the frame areas with poor continuity are separately identified and marked. Then, the interpolation points are calculated frame by frame, and the weighted average of the pixels inserted into the missing frames is generated according to the time axis difference to generate the frame filling result. At the same time, time domain smoothing is used for the local unstable areas of the frame rate to eliminate discontinuous changes. Through adjustment, the smoothness and visual effects of the animation sequence are optimized, and the animation frame repair results are generated.

[0144] A real-time animation capture and repair system, the system comprising:

[0145] The frame change rate analysis module calculates the pixel change rate between consecutive animation frames based on the animation time series data, determines the change threshold between key frames and non-key frames, dynamically adjusts the image acquisition parameters, and obtains the optimal configuration of rendering parameters by comparing and analyzing the frame differences before and after the adjustment;

[0146] The visual element recognition module performs real-time semantic segmentation on the collected animation frames based on the optimized configuration of rendering parameters, identifies key visual elements in the frames frame by frame, and dynamically adjusts the image quality to obtain visual correction records;

[0147] The image quality tuning module compares the quality indicators of the original frame and the repaired frame, including clarity and color accuracy, based on the visual correction record, evaluates the image quality, and adjusts the image restoration parameters according to the evaluation results to obtain frame quality refinement information;

[0148] The timeline alignment module synchronizes the original animation sequence based on the frame quality refinement information, checks the continuity and visual consistency on the timeline through time code comparison, and obtains the timeline calibration result;

[0149] The playback optimization module adjusts the frame interpolation process based on the timeline calibration results, optimizes the smoothness and visual effects of animation playback, and obtains the animation frame repair results.

[0150] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A method for real-time capture and restoration of animation, characterized in that: The following steps are involved: Based on the animation time series data, the pixel change rate between consecutive animation frames is calculated, the change threshold between key frames and non-key frames is determined, the image acquisition parameters are dynamically adjusted according to the change threshold, and the rendering parameter optimization configuration is obtained by comparing and analyzing the frame differences before and after the adjustment; Based on the rendering parameter optimization configuration, the collected animation frames are semantically segmented in real time, key visual elements in the frames are identified frame by frame, including character outlines and backgrounds, and image quality is dynamically adjusted to obtain visual correction records; Based on the visual correction record, quality indicators of the original frame and the repaired frame are compared, including clarity and color accuracy, image quality evaluation is performed, and image repair parameters are adjusted according to the evaluation results to obtain frame quality refinement information; The frame quality refinement information is synchronized with the original animation sequence, and the continuity and visual consistency on the timeline are checked through time code comparison. The frame interpolation process is adjusted according to the timeline calibration result to optimize the smoothness and visual effect of the animation playback, and obtain the animation frame repair result.

2. The method for real-time capture and restoration of animation according to claim 1, characterized in that: The step of determining the change threshold between the key frame and the non-key frame is specifically as follows: Based on the animation time series data, the pixel change rate between consecutive animation frames is calculated using the formula: Get the average rate of change V, where p i，t represents the color value of the i-th pixel at time t, p i，t+1 represents the color value of the i-th pixel at time t+1, and N represents the total number of pixels in the image; The percentile is used to determine the threshold of the average change rate, and key frames and non-key frames are distinguished. If the average change rate exceeds the threshold, the frame is regarded as a key frame, and a change threshold determination result is obtained.

3. The method for real-time capture and restoration of animation according to claim 1, characterized in that: The steps for obtaining the rendering parameter optimization configuration are specifically as follows: Based on the change threshold, adjusting image acquisition parameters, including frame interval and exposure settings, and capturing key frame changes to obtain adjusted acquisition parameter settings; Based on the adjusted acquisition parameter settings, the animation frames are reacquired, and the frame data difference before and after the adjustment is calculated using the formula: Get the frame difference analysis result, where QC is the average pixel difference, p i，new Represents the color value of the i-th pixel after adjustment, p i，old represents the color value of the i-th pixel before adjustment, and N represents the total number of pixels in the image; Based on the frame difference analysis result, the compression ratio and resolution of the rendering parameters are continuously adjusted to match the optimal quality and efficiency balance of animation rendering, thereby obtaining an optimized configuration of the rendering parameters.

4. The method for real-time capture and restoration of animation according to claim 1, characterized in that: The steps of identifying the key visual elements in the frame are specifically as follows: Based on the rendering parameter optimization configuration, animation frame acquisition is performed, and each frame image is verified to meet a predetermined quality standard to obtain an optimized animation frame; The optimized animation frames are subjected to real-time semantic segmentation using the formula: Get the segmentation data set, where SU i represents the segmentation result of the i-th pixel, xu ij is the value of the i-th pixel on the j-th feature, wu j is the weight of the jth feature, M is the total number of features; Based on the segmented data set, key visual elements in the frame, including the outline of a person and the background, are extracted and identified, and the outline of the person and the background are separated to obtain a visual element recognition record.

5. The method for real-time capture and restoration of animation according to claim 1, characterized in that: The steps for obtaining the visual correction record are specifically as follows: Based on the key visual elements, preliminary visual correction is performed on the image, including adjusting light and dark contrast, color saturation and sharpening, to obtain corrected image data; Performing quality assessment on the corrected image data, evaluating the quality difference between the correction effect and the original image, and obtaining a correction effect score; Based on the correction effect score, the image correction parameters are continuously adjusted, and the image processing flow is optimized to obtain a visual correction record.

6. The method for real-time capture and restoration of animation according to claim 1, characterized in that: The step of comparing the quality indicators of the original frame and the repaired frame is specifically as follows: Based on the visual correction record, extracting data of the original frame and the repaired frame, including pixel brightness, color distribution and contrast attributes, to obtain original and repaired frame data sets; Analyzing the original and restored frame data sets, calculating the clarity and color accuracy through peak signal-to-noise ratio and color fidelity index, and obtaining preliminary comparative data of image quality; Based on the preliminary comparative data of the image quality, the formula was used: Evaluate the quality of the image and obtain the image quality index QI, where wo1 and wo2 are weight parameters used to adjust the criticality of PSNR and CFI in quality assessment. PSNR is the peak signal-to-noise ratio and CFI is the color fidelity.

7. The method for real-time capture and restoration of animation according to claim 1, characterized in that: The steps for obtaining the frame quality refinement information are specifically as follows: According to the evaluation results, identifying restoration parameters that cause image quality degradation problems, the problems including insufficient clarity or color distortion, and obtaining degradation problem identification results; According to the degradation problem identification result, image restoration parameters are adjusted, including sharpening degree, color correction setting and contrast adjustment, the image quality is evaluated again, and the adjustment effect is verified to obtain frame quality refinement information.

8. The method for real-time capture and restoration of animation according to claim 1, characterized in that: The steps for obtaining the animation frame repair result are specifically as follows: Synchronizing the frame quality refinement information with the original animation sequence, verifying the correspondence between each repaired frame and the original frame on the time axis through time code comparison, and obtaining calibrated time code data; Analyze the calibrated timecode data to check for continuity and visual consistency on the timeline, using the formula: Get the visual consistency evaluation result, where VCI is the visual consistency index, VF i Represents the visual quality index of the i-th frame, VF i-1 Represents the visual quality index of the i-1th frame, N F Represents the total number of frames, abs(VF i -VF i-1 ) is the absolute value of the visual quality difference between the i-th frame and its previous frame i-1; Based on the visual consistency evaluation result, the frame interpolation process is adjusted to optimize the smoothness and visual effects of the animation playback, including adjusting the frame rate, inserting or deleting the target frame, and obtaining the animation frame repair result.

9. A real-time animation capture and repair system, characterized in that: According to the method for real-time animation capture and restoration according to any one of claims 1 to 8, the system comprises: The frame change rate analysis module calculates the pixel change rate between consecutive animation frames based on the animation time series data, determines the change threshold between key frames and non-key frames, dynamically adjusts the image acquisition parameters, and obtains the optimal configuration of rendering parameters by comparing and analyzing the frame differences before and after the adjustment; The visual element recognition module performs real-time semantic segmentation on the collected animation frames based on the optimized configuration of the rendering parameters, recognizes key visual elements in the frames frame by frame, and dynamically adjusts the image quality to obtain a visual correction record; The image quality tuning module compares the quality indicators of the original frame and the repaired frame based on the visual correction record, including clarity and color accuracy, evaluates the image quality, and adjusts the image repair parameters according to the evaluation results to obtain frame quality refinement information; The timeline alignment module synchronizes the original animation sequence based on the frame quality refinement information, checks the continuity and visual consistency on the timeline through time code comparison, and obtains a timeline calibration result; The playback optimization module adjusts the frame interpolation process based on the time axis calibration result, optimizes the smoothness and visual effect of the animation playback, and obtains the animation frame repair result.

Images