Adaptive Spatial and Motion-Compensation Temporal Filters for Video Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video noise reduction methods fail to effectively address both spatial and temporal noise, leading to residual images or blurring when objects move, as they either overlook temporal correlations or spatial characteristics, resulting in inadequate noise reduction in various environments.
Innovation Solution
The method employs adaptive spatial and motion-compensation temporal filters, using a motion window to calculate pixel means, variances, and determining filter types based on predetermined values, with motion estimation and compensation to filter pixels and remove residual images across frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If motion compensation is applied to reduce temporal noise, then temporal noise reduction is improved, but residual images appear when objects move
Solution Approach 1:
The patent applies different filtering strategies to different regions of the image based on motion characteristics. For stationary regions, temporal filtering is applied to reduce noise. For moving regions, the filter adapts to preserve motion details while reducing noise, preventing residual images. This local differentiation resolves the contradiction by treating temporal noise reduction and motion preservation differently in different spatial locations.
Solution Approach 2:
The patent uses adaptive filtering where the filter parameters dynamically adjust based on motion estimation results. The filter transitions between different modes (temporal filtering for stationary areas, spatial-temporal filtering for moving areas) based on detected motion characteristics. This dynamic adaptation allows the system to reduce temporal noise while preserving motion information, eliminating residual images.
2Reliability
If spatial filtering is applied to reduce noise, then spatial noise is reduced, but image details and edges become blurred
Solution Approach 1:
The patent applies spatial filtering selectively based on local image characteristics. In homogeneous regions, stronger spatial filtering is applied to reduce noise. At edges and detailed regions, the filter strength is reduced or the filter type is changed to preserve sharpness and details. This local adaptation resolves the contradiction between noise reduction and detail preservation.
Solution Approach 2:
The patent uses edge detection and gradient analysis as feedback to control the filtering process. The filter parameters are adjusted based on local gradient magnitude and direction, allowing the system to automatically preserve edges while filtering noise in flat regions. This feedback mechanism ensures that image details are maintained while achieving effective noise reduction.
3Reliability
If strong filtering is applied to reduce noise, then noise reduction is improved, but image quality and sharpness deteriorate
Solution Approach 1:
The patent implements adaptive filtering where the filter strength is locally adjusted based on noise characteristics and image content. In noisy homogeneous regions, stronger filtering is applied. In regions with important features or low noise, filtering is reduced to preserve quality. This local quality approach resolves the contradiction by optimizing the trade-off between noise reduction and quality preservation in different regions.
Solution Approach 2:
The patent dynamically changes filter parameters (filter size, kernel type, weighting factors) based on local image statistics and noise characteristics. The system adjusts parameters to achieve optimal noise reduction while maintaining image quality, rather than applying fixed strong filtering throughout. This parameter adaptation resolves the contradiction by making filtering intensity context-dependent.
4Reliability
If temporal filtering without motion compensation is used, then temporal noise is reduced, but moving objects produce residual images and fuzzy phenomena
Solution Approach 1:
The patent transitions from static temporal filtering to dynamic motion-compensated filtering. Motion estimation provides dynamic information about object movement, which is used to adjust the filtering process. The filter compensates for motion by aligning temporal samples before averaging, preventing residual images while maintaining noise reduction benefits. This dynamic approach resolves the contradiction between temporal noise reduction and motion preservation.
Solution Approach 2:
The patent introduces motion compensation as an intermediary step between temporal sampling and filtering. The motion compensation process creates a corrected temporal alignment that accounts for object movement, serving as a mediator that enables both noise reduction and motion preservation. This intermediary mechanism resolves the contradiction by eliminating the source of residual images while maintaining temporal filtering benefits.
Data Source
AI summary
A method includes calculating a mean of a plurality of pixels of a motion window, calculating a pixel amount of pixels similar to a center pixel, calculating a variance of the pixels, determining whether a difference between the center pixel and the mean is greater than a first predetermined value, determining whether the pixel amount similar to the center pixel is greater than a second predetermined value if the difference between the center pixel and the mean is not greater than the first predetermined value, determining whether the variance is smaller than a threshold value if the pixel amount similar to the center pixel is not greater than the second predetermined value, and filtering the center pixel according to a result of determining whether the variance is smaller than the threshold value. Finally, temporal weighted mean filters involving motion estimation are used for motion compensation in images after spatial filtering.