Temporal Intra Mode Derivation from Displaced Collocated Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding and decoding technologies face challenges in efficiently deriving intra modes, leading to increased signaling overhead and reduced compression efficiency due to the need to signal intra prediction mode indices explicitly.
Innovation Solution
The proposed method involves determining a motion vector pointing to a reference frame, deriving a displaced collocated block, and conditionally adding intra mode information to a most probable mode list for improved decoder-side intra mode derivation, utilizing techniques like Template-based Intra Mode Derivation (TIMD) and Subblock-Based Temporal Motion Vector Prediction (SbTMVP) to enhance coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If intra prediction mode indices are signaled explicitly, then decoding accuracy is improved, but signaling overhead increases
Solution Approach 1:
The decoder derives intra prediction modes autonomously from the displaced collocated block without requiring explicit mode indices from the encoder. The displaced block's intra modes are directly reused or slightly modified to generate the current block's prediction modes, enabling the decoder to self-determine modes without additional signaling.
Solution Approach 2:
Intra prediction modes from the displaced collocated block are copied and reused for the current block. Instead of signaling new mode indices, the decoder copies mode information from the temporally corresponding block in the reference frame, eliminating the need for explicit mode transmission while maintaining prediction accuracy.
2Measurement precision
If more intra mode information is signaled, then prediction accuracy is improved, but compression efficiency deteriorates
Solution Approach 1:
The decoder self-determines intra prediction modes by deriving them from the displaced collocated block's mode information. This eliminates the need for the encoder to signal multiple candidate modes or mode indices, reducing bitstream overhead while maintaining accurate prediction through temporal reuse of mode data.
Solution Approach 2:
Instead of signaling redundant mode information, the method discards explicit mode signaling and recovers mode data by reusing it from the displaced collocated block. This trade-off reduces signaling overhead while preserving prediction accuracy through intelligent reuse of temporal information.
Data Source
AI summary
Intra prediction modes are derived in video encoders and decoders using one of several embodiments. In at least one embodiment, an intra prediction mode is derived when little or no spatial information is available. In another embodiment, motion information is extracted and a corresponding block is obtained from a displaced collocated block which is used to determine an intra prediction mode for encoding/decoding a video block or sub-block. In another embodiment, a global motion model is computed at an encoder and sent in a slice header or a picture header to a corresponding decoder. In at least one embodiment, a different reference is used for a reference frame. In at least one embodiment, an index is signaled to indicate the intra mode, reference frame, or motion model is used for prediction.