Last Significant Coefficient Signaling in Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding schemes fail to accurately signal the position of the last significant coefficient in a block after transform and quantization, as they ignore zeroing processes that can alter the position of non-zero coefficients, leading to inefficiencies in compression and decoding.
Innovation Solution
A method and device that determine and signal the position of the last significant coefficient based on the zeroing process applied to the block, considering the type of transform process used, allowing for more accurate encoding and decoding by accounting for zeroing effects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the position of the last significant coefficient is signaled with respect to the top-left or bottom-right corner of the sub-block, then the signaling is simple and based on known signal properties, but the accuracy is reduced because zeroing processes that alter coefficient positions are ignored
Solution Approach 1:
The patent applies preliminary action by determining whether a zeroing process has been applied to the block before signaling the last significant coefficient position. The encoder/decoder checks if zeroing was performed (based on transform type or explicit flags) and adjusts the reference position accordingly in advance, ensuring accurate position signaling without adding complex real-time processing
Solution Approach 2:
The patent implements dynamics by making the reference position for signaling the last significant coefficient adaptive rather than fixed. The reference position changes dynamically based on the transform type applied (e.g., DST-II uses bottom-right corner, DCT-II uses top-left corner) or explicit zeroing flags, allowing the system to optimize positioning accuracy for different coding scenarios
2Productivity
If transform and quantization are applied to exploit spatial and temporal redundancies, then compression efficiency is improved, but the position of non-zero coefficients is altered by subsequent zeroing processes, leading to inaccurate last significant coefficient signaling
Solution Approach 1:
The patent applies feedback by using information about the transform type and zeroing process applied to the block to adjust how the last significant coefficient position is signaled. The encoder determines whether zeroing was applied (based on transform type or flags) and feeds this information into the positioning logic, ensuring the decoder can accurately reconstruct the position despite zeroing effects
Solution Approach 2:
The patent implements parameter changes by modifying the reference position parameter used for signaling based on the transform type and zeroing process. Instead of always using top-left or bottom-right corner as reference, the system changes the reference position parameter adaptively (e.g., using bottom-right for DST-II, top-left for DCT-II) to maintain accuracy after transform and quantization
Data Source
AI summary
A method comprising: determining (701) that a zeroing process was applied to a block of transform coefficients; and, decoding (702) a position of a last significant coefficient of the block in scanning order with respect to a position in the block depending on the applied zeroing process.