Recursive Scene Segmentation for HDR Video Metadata Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In cloud-based video coding architectures, existing techniques for processing HDR video often require frame-by-frame updates of reshaping metadata, leading to significant overhead, especially at low bit rates, due to the need for scene-based processing across multiple computing nodes.
Innovation Solution
The proposed solution involves segmenting video frames into scenes and sub-scenes, using a recursive scene splitting algorithm to minimize the need for reshaping function updates, thereby reducing metadata overhead by applying scene-based forward and backward reshaping functions to maintain temporal consistency across nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If frame-by-frame updates of reshaping metadata are used, then video quality is maintained, but metadata overhead increases significantly
Solution Approach 1:
The video sequence is divided into multiple scenes, and within each scene, frames are grouped into segments that share common reshaping characteristics. This segmentation allows reshaping metadata to be updated at the segment level rather than frame-by-frame, reducing metadata overhead while maintaining video quality through localized updates only when scene characteristics change.
2Productivity
If scene-based processing is applied across multiple computing nodes, then processing efficiency is improved, but metadata overhead increases
Solution Approach 1:
The patent segments video sequences into scenes that can be independently processed by different computing nodes. Each node processes a specific scene and generates reshaping metadata only for its assigned segment, eliminating the need for global frame-by-frame metadata updates across all nodes and reducing overall metadata overhead while maintaining parallel processing efficiency.
Solution Approach 2:
The patent applies local quality by generating reshaping metadata with local validity scope limited to specific scenes or segments rather than globally across the entire video sequence. Each computing node generates metadata locally for its assigned scene, reducing redundant metadata transmission and processing across the distributed system while maintaining processing efficiency.
3Quantity of substance
If recursive scene splitting is implemented, then metadata overhead is reduced, but computational complexity increases
Solution Approach 1:
The patent applies preliminary action by performing recursive scene splitting and segment identification during the initial video analysis phase. The video sequence is pre-segmented into scenes and segments with assigned validity ranges for reshaping metadata before actual encoding begins. This preliminary segmentation reduces the computational burden during the main encoding process and minimizes metadata overhead by establishing segment boundaries in advance.
Data Source
AI summary
In a cloud-based system for encoding high dynamic range (HDR) video, each node receives a video segment and bumper frames. Each segment is subdivided into primary scenes and secondary scenes to derive scene-based forward reshaping functions that minimize the amount of reshaping-related metadata when coding the video segment, while maintaining temporal continuity among scenes processed by multiple nodes. Methods to generate scene-based forward and backward reshaping functions to optimize video coding and improve the coding efficiency of reshaping-related metadata are also examined.