Temporal Noise Reduction via Channel-Specific Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional temporal noise reduction methods in video processing are inaccurate and degrading image quality due to oversimplified motion detection and reliance on single-channel information for filtering decisions, leading to inadequate filtering of other channels.
Innovation Solution
A method involving motion detection and pixel difference calculations to determine weighting values for temporal filtering, with separate filtering modules for each channel to accurately assess and reduce noise in video data streams, ensuring each channel is filtered based on its own information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional temporal noise filtering is performed using a single predetermined threshold value for motion detection, then the filtering process is simple and fast, but the accuracy of noise reduction deteriorates due to oversimplified motion detection and single-channel information reference
Solution Approach 1:
The patent divides the filtering process into multiple independent channel processing units (Y channel, U channel, V channel), where each channel performs its own motion detection and filtering operations separately rather than using a single unified process. This segmentation allows each channel to be optimized independently for both speed and accuracy.
Solution Approach 2:
The patent replaces the static predetermined threshold with dynamic threshold values that are adaptively determined for each channel based on actual image characteristics. The controller dynamically adjusts threshold values according to the specific content and noise characteristics of each channel, enabling accurate noise reduction while maintaining processing efficiency.
2Device complexity
If motion detection uses a predetermined threshold value to determine image motion, then the control logic is simple, but the filtering accuracy deteriorates because the determination is crude and does not account for different channel characteristics
Solution Approach 1:
The patent applies different threshold values and filtering parameters to different channels (Y, U, V) based on their specific characteristics. Each channel has its own optimized threshold and filtering strength, allowing accurate motion detection adapted to the unique properties of each color channel rather than using a one-size-fits-all approach.
Solution Approach 2:
The patent changes the threshold parameter from a fixed predetermined value to dynamically adjustable values that can be independently optimized for each channel. The controller modifies threshold parameters based on actual image content and noise characteristics, enabling accurate motion detection without excessive complexity.
3Ease of manufacture
If only a single channel (e.g., Y channel) is used for motion detection to control filtering of all channels, then hardware costs and computational resources are minimized, but image quality deteriorates because other channels are inadequately filtered
Solution Approach 1:
The patent implements separate filtering modules for Y, U, and V channels, where each channel has its own motion detection and filtering processing path. This segmentation ensures that each channel is filtered based on its own characteristics rather than relying on another channel's motion detection results, maintaining image quality without excessive hardware complexity.
Solution Approach 2:
The patent creates a universal filtering architecture where the same filtering mechanism and control logic can be applied to multiple channels (Y, U, V) simultaneously. Each channel processing unit is designed to be functionally equivalent but independently controllable, allowing the system to maintain reliability across all channels while using a standardized approach that doesn't significantly increase hardware complexity.
Data Source
AI summary
Disclosed is a method for reducing temporal noise, comprising: performing motion detection on frames of a video data stream; calculating a pixel difference between pixels of frames in the video data stream to generate at least a pixel difference value; determining a set of weighting value for temporal filtering according to a result of the motion detection and a result of the pixel difference value calculation; and performing temporal filtering on frames in the video data stream according to the weighting values.