Image Tile Parallel Processing Encoding Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image encoding and decoding methods, such as H.264 and HEVC, face challenges in achieving high-speed parallel processing due to complexities in controlling parallel processing tasks, especially with large images like super high-definition screens, where dividing the image into tiles for multiple processor nodes is not feasible without reducing encoding efficiency.
Innovation Solution
The proposed solution involves an image encoding and decoding apparatus that divides a picture into rectangular tiles, allowing for parallel processing by using a combination of tile dividing units and encoding units, with a control unit selecting appropriate encoding methods to enable high-speed encoding-decoding processes. This includes a decoding unit that decodes images based on a code indicating whether parallel processing is enabled for each tile, and an encoding unit that generates codes capable of sequential or parallel decoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the picture is divided into multiple tiles for parallel processing, then processing speed is improved, but control complexity increases
Solution Approach 1:
The picture is divided into multiple rectangular tiles that can be processed independently in parallel. Each tile is further divided into blocks and lines, allowing hierarchical parallel processing. This segmentation enables multiple processor nodes to work simultaneously on different tiles, improving processing speed while maintaining manageable control complexity through structured division.
Solution Approach 2:
The patent introduces dynamic control mechanisms where the parallel processing configuration can be adjusted based on picture size and processing requirements. For small pictures, sequential processing is used; for large pictures, parallel processing across multiple tiles is enabled. This dynamic adaptation optimizes processing speed while avoiding unnecessary control complexity for small-scale processing.
2Productivity
If statistical information is reset for parallel encoding, then parallel processing is enabled, but encoding efficiency is reduced
Solution Approach 1:
The picture is divided into multiple slices, with each slice containing multiple tiles. Statistical information (context models) is reset at slice boundaries rather than tile boundaries, allowing parallel processing within slices while maintaining encoding efficiency through longer contexts. This hierarchical segmentation enables parallel processing at the tile level while preserving statistical efficiency at the slice level.
Solution Approach 2:
The slice structure acts as an intermediary between the tile-level parallel processing and the statistical model management. By introducing slices as intermediate units, the system can reset statistical information less frequently (at slice boundaries) while still enabling parallel processing at the finer tile level, thus mediating between parallelism requirements and encoding efficiency.
3Area of stationary object
If the picture size is large (super high-definition), then parallel processing becomes necessary, but control complexity increases significantly
Solution Approach 1:
Large pictures are divided into multiple tiles arranged in a grid pattern, with each tile further divided into slices and blocks. This hierarchical segmentation allows large pictures to be processed in manageable units, distributing the control complexity across multiple levels of organization rather than requiring monolithic control of the entire large picture.
Solution Approach 2:
The patent introduces a hierarchical dimension to parallel processing control, organizing tiles into slices and slices into pictures. This adds temporal and organizational dimensions to the spatial division, allowing complex large-picture processing to be managed through multi-level hierarchical control rather than flat single-level control.
Data Source
AI summary
An image encoding format includes a profile indicating a combination of processes which an image decoding apparatus is capable of decoding and a level indicating a range of a parameter which the image decoding apparatus is capable of decoding. The image decoding apparatus includes a decoding unit that decodes the image encoding format on the basis of a code that indicates whether parallel processing of the decoding processes in each tile is enabled depending on the profile.