Video Encoder Block Partitioning with Conditional Parameter Signaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image and video encoding methods face challenges in efficiently partitioning blocks for encoding and decoding processes, particularly with non-standard sizes, leading to suboptimal performance and efficiency.
Innovation Solution
An encoder and decoder system that utilizes a set of block partition modes, including a first partition mode defining the direction and number of partitions for a block, and a second partition mode for a center block with a fixed number of partitions, using a flag to indicate horizontal or vertical partitioning, without specifying the total number of partitions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a complete block partition mode is specified for every block including the total number of partitions, then the partitioning flexibility is improved, but the bitstream complexity and encoding overhead increase
Solution Approach 1:
The patent applies partial action by specifying block partition modes only for blocks that require partitioning, while omitting the total number of partitions parameter for blocks where the partition count is predetermined (e.g., when partition direction is horizontal/vertical and total partitions is 3). This reduces bitstream complexity while maintaining partitioning flexibility for cases where it is needed.
Solution Approach 2:
The patent changes the parameter specification strategy by introducing conditional inclusion of parameters based on block characteristics. When a block is identified as a center block with specific partition direction and total partitions, the system changes from specifying all parameters to specifying only the necessary partition direction parameter, thereby reducing overhead while preserving adaptability.
2Measurement precision
If the total number of partitions is specified for every block, then the encoding precision is improved, but the encoding efficiency decreases
Solution Approach 1:
The patent extracts the redundant information (total number of partitions parameter) from the bitstream for blocks where it can be determined from the partition direction and block type. This extraction reduces the amount of data to be encoded and decoded, thereby improving encoding efficiency while maintaining sufficient precision through the retained partition direction information.
Solution Approach 2:
The system enables self-service by allowing the decoder to automatically determine the total number of partitions based on the partition direction and block characteristics without requiring explicit parameter transmission. This self-determination mechanism improves encoding efficiency by reducing bitstream size while maintaining encoding precision through logical inference.
Data Source
AI summary
An encoder partitions into blocks using a set of block partition modes. The set of block partition modes includes a first partition mode for partitioning a first block, and a second block partition mode for partitioning a second block which is one of blocks obtained after the first block is partitioned. When the number of partitions of the first block partition mode is three, the second block is a center block among the blocks obtained after partitioning the first block, and the partition direction of the second block partition mode is same as the partition direction of the first block partition mode, the second block partition mode indicates that the number of partitions is only three. A parameter for identifying the second block partition mode includes a first flag indicating a horizontal or vertical partition direction, and does not include a second flag indicating the number of partitions.