Video Subpicture Encoding for CTB Boundary Splitting Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Decoders face challenges in inferring the splitting of coding tree blocks (CTBs) at the right or bottom boundaries of subpictures when their size is not a multiple of the CTB size, leading to incomplete encoding and decoding failures during the merging of subpictures from different video bitstreams.
Innovation Solution
Encoding methods are introduced to include information in the bitstream indicating that subpictures have a size that is a multiple of coding tree blocks and are independently decodable, along with conformance windows, allowing decoders to infer the required splitting of CTBs at the boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the subpicture size is not a multiple of the coding tree block size, then the subpicture can be more flexible in size adaptation, but the decoder cannot infer the splitting of rightmost or bottommost coding blocks leading to decoding failures
Solution Approach 1:
The encoder performs preliminary action by explicitly signaling the subpicture width and height in the bitstream before decoding occurs. This allows the decoder to have advance knowledge of the exact subpicture dimensions, enabling it to correctly infer the splitting of coding tree blocks at the rightmost or bottommost boundaries even when the subpicture size is not a multiple of the CTB size, thereby preventing decoding failures
Solution Approach 2:
The subpicture width and height parameters act as intermediaries between the encoder and decoder. These explicitly signaled dimensions serve as a bridge that allows the decoder to correctly interpret the bitstream structure and perform accurate CTB splitting inference, resolving the information asymmetry that would otherwise lead to decoding failures
2Adaptability or versatility
If the subpicture is moved from the right or bottom boundary to other locations, then the subpicture can be rearranged flexibly, but the decoder fails to identify missing coding blocks
Solution Approach 1:
The encoder performs preliminary action by explicitly signaling the subpicture width and height before any rearrangement or displacement of the subpicture occurs. This advance information allows the decoder to correctly identify the spatial extent of the subpicture regardless of its position in the picture, enabling accurate identification of coding blocks even when the subpicture is moved from the right or bottom boundary to other locations
Solution Approach 2:
The explicit subpicture width and height parameters provide feedback information to the decoder about the subpicture's spatial dimensions. This feedback mechanism allows the decoder to verify and correctly identify the location and extent of coding blocks within the subpicture, preventing information loss during subpicture rearrangement operations
Data Source
AI summary
An encoding method comprising the encoding of information allowing the decoder to infer the splitting of CTBs located at the right, respectively the bottom of a subpicture, which width, respectively height, is not a multiple of the size of the CTBs, when the subpicture is not located at the right, respectively the bottom, of the image. Corresponding decoding methods for the generated bitstream are also proposed.