Subpicture Inter Coding with Matching CTU Size
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards, such as HEVC and VVC, face challenges in efficiently encoding and decoding video data when pictures are partitioned into subpictures, particularly in maintaining consistent Coding Tree Unit (CTU) size and subpicture layout across reference and current pictures.
Innovation Solution
The proposed method involves partitioning a current picture into subpictures and determining valid reference pictures for inter coding, where each reference picture must have the same CTU size and subpicture layout as the current picture, unless it is an Inter Layer Reference Picture (ILRP) containing one subpicture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If pictures are partitioned into subpictures with different CTU sizes, then coding flexibility and adaptability to local characteristics are improved, but inter coding compatibility and decoding complexity increase
Solution Approach 1:
The picture is divided into multiple subpictures, each with its own CTU size configuration. This segmentation allows different regions to use optimal CTU sizes for their local characteristics while maintaining independent decoding paths, thus improving coding flexibility without overwhelming the decoder with a single complex configuration.
Solution Approach 2:
The patent enables dynamic selection of reference pictures with matching CTU sizes for each subpicture. This dynamic adaptation allows the system to adjust reference picture selection based on the current subpicture's CTU configuration, improving inter coding compatibility while maintaining coding flexibility.
2Productivity
If reference pictures with different CTU sizes are used for inter coding, then coding efficiency is improved, but bitstream conformance and decoding reliability deteriorate
Solution Approach 1:
The patent introduces syntax elements to signal CTU size information for reference pictures and implements constraints that ensure reference pictures have compatible CTU sizes. This parameter control mechanism maintains bitstream conformance while still allowing efficient inter coding by selecting appropriate reference pictures.
Solution Approach 2:
The decoder uses signaled CTU size information from reference pictures to validate and adjust its decoding process. This feedback mechanism ensures that only compatible reference pictures are used, maintaining bitstream conformance while enabling efficient inter coding when conditions are met.
3Adaptability or versatility
If subpicture layout varies between reference and current pictures, then adaptability to different picture contents is improved, but inter coding accuracy and prediction quality worsen
Solution Approach 1:
The patent allows different subpicture layouts for different pictures while ensuring that within each picture, subpictures maintain consistent CTU size and layout. This local consistency within pictures combined with global adaptability across pictures achieves both picture content adaptability and prediction quality.
Solution Approach 2:
The system dynamically selects reference pictures based on layout compatibility requirements. When a reference picture has a compatible subpicture layout, it is selected for inter coding; otherwise, alternative reference pictures or intra coding is used. This dynamic selection maintains prediction quality while preserving overall adaptability.
Data Source
AI summary
Video processing methods and apparatuses include receiving input video data associated with a current picture, signaling or parsing a first syntax element, wherein the first syntax element indicates whether to infer or predict subpicture information of a current layer of the current picture from a target layer; signaling or parsing a second syntax element to determine information about the target layer for obtaining the subpicture information of the current layer of the current picture, and encoding or decoding the current layer of the current picture using the subpicture information of the current layer of the current picture.