Decoder-Derived Intra Prediction for Non-Square Video Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards like VVC face challenges in efficiently handling non-square blocks and large transform sizes, leading to increased data and complexity, particularly in hardware decoders, due to the use of ternary and binary tree partitions and wide-angle intra prediction modes.
Innovation Solution
The method employs decoder-side intra mode derivation (DIMD) and template-based intra mode derivation (TIMD) to determine prediction directions for current blocks, using statistics or histograms, and applies block differential pulse coded modulation (BDPCM) and geometric partitioning mode (GPM) to optimize intra and inter coding, reducing data and complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If ternary and binary tree partitions are used for non-square blocks, then coding flexibility and adaptability are improved, but data quantity and device complexity increase
Solution Approach 1:
The patent extracts only the essential partitioning information needed for decoding by using simplified flags to indicate partition types (ternary, binary, or none) rather than transmitting complete partitioning data structures. This reduces the data overhead while maintaining the ability to handle non-square blocks effectively.
Solution Approach 2:
Instead of explicitly signaling all partitioning details from encoder to decoder, the patent inverts the approach by having the decoder derive partitioning information implicitly from the coded data and context, reducing the need for explicit signaling and thereby lowering device complexity.
2Measurement precision
If wide-angle intra prediction modes are used, then prediction accuracy is improved, but data quantity and complexity increase
Solution Approach 1:
The patent changes the parameter representation by using compact flags to indicate wide-angle prediction mode selection rather than transmitting full mode information. This allows the decoder to reconstruct the prediction modes with reduced data overhead while maintaining prediction accuracy.
Solution Approach 2:
The patent uses lightweight, easily computable prediction mode indicators that can be quickly processed and discarded after use, replacing complex, long-lived data structures. This reduces both memory usage and processing complexity while maintaining prediction quality.
3Measurement precision
If explicit mode signaling is used, then decoding accuracy is improved, but bit rate increases
Solution Approach 1:
The patent transforms the signaling parameters from explicit mode indicators to implicit derivation mechanisms, where the decoder reconstructs mode information from compact flags and context. This maintains decoding accuracy while significantly reducing the bit rate required for transmission.
Solution Approach 2:
The patent introduces intermediary flags and context information that mediate between the encoded data and the final decoded modes. These intermediaries carry essential information in a compressed form, allowing accurate reconstruction without transmitting full mode details explicitly.
Data Source
AI summary
Methods and apparatus for video coding are disclosed. One method determines a prediction direction between vertical prediction and horizontal prediction for the current block based on a template of the current block or based on decoder side intra mode derivation (DIMD) using statistics or histogram of angle field derived from the template of the current block. According to another method, the current block is partitioned into a first region and a second region according to a region split. The first region is encoded or decoded based on inter coding. The second region is encoded or decoded according to intra coding. At least a part of region-spit parameters, a part of inter coding parameters, or a part of intra coding parameters are determined based on a template of the current block or based on DIMD using statistics or histogram of angle field derived from the template of the current block.