Intra Image Coding Using Residual Prediction Without Mode Signaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image coding methods require inefficient bit allocation for residual signals and intra prediction modes, leading to suboptimal coding efficiency.
Innovation Solution
Applying residual prediction to intra prediction and determining intra prediction modes on both encoder and decoder sides without explicit signaling, reducing the number of bits required for coding residual signals and intra prediction modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If residual prediction is applied to intra prediction, then the number of bits necessary for coding residual signal is reduced, but device complexity increases
Solution Approach 1:
The patent merges residual prediction with intra prediction by determining a reference block's residual signal and applying it to predict the current block's residual signal. This combination reduces the number of bits needed for coding by utilizing already-reconstructed residual information from reference blocks, thereby resolving the contradiction between reducing bit requirements and managing processing complexity.
Solution Approach 2:
The patent performs preliminary action by pre-determining reference blocks and their residual signals before processing the current block. The reference block's residual signal is calculated in advance using previously reconstructed signals, enabling efficient prediction of the current block without requiring complex real-time analysis, thus reducing coding bits while controlling complexity.
2Loss of information
If intra prediction mode is determined without explicit signaling, then the number of bits for signaling is reduced, but reliability of prediction mode alignment between encoder and decoder deteriorates
Solution Approach 1:
The patent applies self-service by enabling the encoder and decoder to independently determine the same intra prediction mode without explicit signaling. The prediction mode is derived from the block's position, size, and neighboring block information, which are already available at both ends, eliminating the need for signaling while ensuring reliable alignment through shared contextual knowledge.
Solution Approach 2:
The patent implements feedback mechanisms where the encoder and decoder continuously reference previously determined prediction modes and block characteristics. This feedback loop ensures that both sides arrive at the same prediction mode determination, maintaining reliability without requiring explicit mode signaling, thus reducing bit requirements while preserving synchronization.
Data Source
AI summary
A method, performed by an apparatus, of decoding a bitstream, includes obtaining a residual difference signal of a current block by processing the bitstream, determining a reference block corresponding to the current block, in a current picture including the current block, determining a residual signal of the reference block, based on a reconstructed signal of the reference block and a prediction signal of the reference block, and determining a reconstructed signal of the current block, based on the residual difference signal of the current block, the residual signal of the reference block, and a prediction signal of the current block. The prediction signal of the reference block is determined based on at least one reference pixel included in a neighboring block of the reference block in the current picture, and the prediction signal of the current block is determined based on at least one reference pixel included in a neighboring block of the current block in the current picture.