Adaptive Clipping in Video Coding Residual Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression systems face challenges in efficiently clipping video data to avoid data overflow or underflow, particularly when high values exceed the representation range, leading to potential errors in video encoding and decoding processes.
Innovation Solution
The method involves adaptive clipping, where low-pass filtered values are used to replace elements in residual blocks if they exceed clipping bounds, and encoding and decoding are performed with these modified blocks to form reconstructed blocks within specified clipping ranges, allowing for efficient transformation and quantization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If video data is clipped to avoid data overflow or underflow, then data representation reliability is improved, but distortion increases when values are forced into the clipping range
Solution Approach 1:
The patent implements dynamic clipping by adjusting clipping thresholds based on local image characteristics and gradient information. Instead of using fixed clipping bounds, the system adaptively determines clipping values for different regions, allowing the clipping operation to respond to local content variations and minimize distortion while maintaining reliability.
Solution Approach 2:
The patent changes the parameters used for clipping by introducing gradient-based adaptive thresholds. The clipping bounds are modified based on local gradient magnitude and direction, transforming the static clipping process into a dynamic one that preserves more image information and reduces artifacts.
2Productivity
If adaptive clipping with low-pass filtering is applied to residual blocks, then compression efficiency is improved, but computational complexity increases
Solution Approach 1:
The patent applies low-pass filtering to residual blocks before the clipping operation. This preliminary action smooths the residual values, reducing high-frequency variations that would otherwise cause clipping artifacts. By preparing the residual block in advance, the subsequent clipping operation becomes more effective and requires less complex post-processing.
Solution Approach 2:
The low-pass filter acts as an intermediary between the residual block and the clipping operation. It mediates the transition by smoothing extreme values before they reach the clipping stage, reducing the need for aggressive clipping and thereby improving compression efficiency without requiring overly complex algorithms.
3Reliability
If clipping is applied to prevent data inversion, then data integrity is improved, but loss of information occurs when high values are treated as low values
Solution Approach 1:
The patent dynamically adjusts clipping parameters based on local image characteristics, using gradient information to determine appropriate clipping thresholds. This adaptive approach changes the clipping behavior from a fixed parameter operation to a dynamic one that preserves more information by only clipping when necessary and using context-aware thresholds.
Solution Approach 2:
The patent applies different clipping strategies to different regions of the image based on local characteristics. By analyzing gradient magnitude and direction in each region, the system applies clipping selectively where needed while preserving information in regions where clipping would be harmful, thus maintaining data integrity while minimizing information loss.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Clipping may be applied during various operations at an encoder or decoder. The lower and upper bounds for clipping can be differentially encoded with respect to predictors. Fixed predictors can be used, for example, 0 and 2bitdepth for the lower and upper bounds, respectively. Adaptive predictors can also be used. Adaptive predictors can be derived from clipping bounds in previously encoded or decoded pictures, or explicitly signaled in a bitstream. Whether to encode the clipping bounds can be determined based on the number of pixels that have values close to the clipping bounds. Additionally, taking advantage of the clipping operation, the prediction residuals can be smoothed, while the distortion for the block may not necessarily increase.