Video Noise Reduction via Motion Trajectory Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern video compression systems face challenges in reducing noise, particularly at low bit rates, where coding artifacts lead to inaccurate predictions and increased noise variance, especially when using predictive error coding and post-filtering techniques.
Innovation Solution
The method involves constructing and combining motion-shifted versions of pixel amplitudes along individual motion trajectories using a weighting function to form processed pixel amplitudes, which are then included in the video sequence, allowing for efficient noise reduction through recursive accumulation and adaptive allocation of motion trajectory lengths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If lossy compression algorithms are used to adjust bit rate to channel capacity, then bandwidth efficiency is improved, but visible coding artifacts and noise increase in the video sequence
Solution Approach 1:
The video sequence is divided into blocks of pixels, and motion trajectories are constructed for individual blocks rather than entire frames. This segmentation allows selective filtering of noise while preserving important coding artifacts that represent actual scene changes, thereby reducing visible noise without significantly increasing bit rate
Solution Approach 2:
The patent extends the filtering approach from spatial domain to temporal domain by constructing motion trajectories across multiple frames. By averaging pixel amplitudes along the temporal dimension (motion trajectory) rather than only in spatial neighborhood, the method reduces noise more effectively while maintaining bandwidth efficiency
2Measurement precision
If motion trajectories are constructed for all pixels in a block, then noise reduction accuracy is improved, but computational complexity increases
Solution Approach 1:
Different blocks in the video sequence are assigned different motion trajectory lengths based on their local characteristics. Blocks with coherent motion (e.g., background) use longer trajectories for better noise reduction, while blocks with complex motion (e.g., foreground objects) use shorter trajectories. This local adaptation improves noise reduction accuracy where needed while reducing unnecessary computations in other areas
Solution Approach 2:
Instead of constructing motion trajectories for all pixels uniformly, the patent applies trajectory construction selectively to blocks that benefit most from noise reduction. The motion trajectory length is adapted per block, using partial action (shorter trajectories) where sufficient and excessive action (longer trajectories) where needed, optimizing the balance between accuracy and complexity
Data Source
AI summary
An embodiment of the invention relates to a method for processing the pixel amplitude of at least one block image pixel contained in a video sequence, the method comprising the steps of: constructing an individual motion trajectory comprising motion-shifted versions of said block image pixel over a multiplicity of neighboring frames; and combining the pixel amplitudes of the motion-shifted versions of said block image pixel along the individual motion trajectory using a weighting function, to form a processed pixel amplitude of said image pixel.