Scalable Video Coding with Content-Adaptive Color Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing scalable video coding systems inefficiently handle High Dynamic Range (HDR) and Wide Color Gamut (WCG) video by using static prediction methods unrelated to video content, requiring separate encoding and transmission of non-HDR and HDR bitstreams, which is time-consuming and inefficient.
Innovation Solution
A method where a video encoder performs content-adaptive color mapping operations on sub-pictures to convert values into a narrower color space for the base layer and uses reverse color mapping operations based on received parameters to reconstruct reference frames, allowing for dynamic prediction in the enhancement layer, and sends these parameters to the decoder for adaptive decoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If separate encoding and transmission of non-HDR and HDR bitstreams is performed, then compatibility with various display devices is ensured, but encoding time and transmission efficiency deteriorate
Solution Approach 1:
The patent merges non-HDR and HDR bitstreams into a single scalable bitstream structure where the base layer contains downsampled non-HDR values and the enhancement layer contains HDR values. This combining approach allows a single encoding operation to produce content compatible with both standard and HDR displays, eliminating the need for separate encoding processes while maintaining broad device compatibility.
Solution Approach 2:
The patent segments the video content into different layers with distinct characteristics: a base layer containing downsampled values in a first color space suitable for non-HDR displays, and an enhancement layer containing full-resolution HDR values in a second color space. This segmentation allows the encoding process to handle different quality requirements separately while producing a unified bitstream that can be adaptively decoded based on display capabilities.
2Device complexity
If static prediction method unrelated to video content is used, then decoding complexity is reduced, but prediction accuracy and video quality deteriorate
Solution Approach 1:
The patent introduces dynamic prediction by performing reverse color mapping operations on base layer values to generate reference pictures for predicting enhancement layer values. The prediction process adapts to local video content characteristics through content-adaptive color mapping, allowing the system to adjust prediction behavior based on actual picture content rather than using a fixed static method, thereby improving prediction accuracy while maintaining reasonable decoding complexity.
Solution Approach 2:
The patent changes the color space parameters dynamically during prediction by applying reverse color mapping operations that convert base layer values from a first color space to match the second color space of the enhancement layer. This parameter transformation enables accurate cross-layer prediction by ensuring color space consistency between reference and current pictures, directly improving prediction accuracy without significantly increasing decoding complexity.
3Ease of manufacture
If color mapping operation is applied uniformly to all pictures, then encoding simplicity is maintained, but encoding precision and video quality deteriorate
Solution Approach 1:
The patent applies local quality by performing content-adaptive color mapping operations on individual sub-pictures rather than uniformly across the entire picture. The encoder analyzes local content characteristics and applies appropriate color mapping parameters to each sub-picture, allowing different regions with different content characteristics to be encoded with optimized color space transformations. This approach significantly improves encoding precision while maintaining reasonable encoding simplicity through automated content-based parameter selection.
Solution Approach 2:
The patent changes color mapping parameters adaptively based on local picture content by analyzing characteristics of each sub-picture and selecting appropriate mapping operations. Instead of using fixed uniform parameters, the system dynamically adjusts color space conversion parameters to match local content requirements, thereby improving encoding precision while keeping the encoding process manageable through systematic parameter adaptation rather than complex manual configuration.
Data Source
AI summary
A method is provided for encoding a digital video to provide for improved color mapping. The digital video has values in a first color space, and the method includes performing a color mapping operation on values in each sub-picture to convert the values in the first color space to values in a second, narrower, color space, wherein the color mapping operation is adapted based on the content of each sub-picture, encoding the values in the second color space into a base layer, performing a reverse color mapping operation on decoded values from the base layer in the second color space in each sub-picture to generate a reconstructed reference frame having values in the first color space, encoding values in the first color space into an enhancement layer based at least in part on the reconstructed reference frame, combining the base layer and the enhancement layer into a bitstream, sending the bitstream to a decoder, and sending one or more parameters to the decoder that describe the adaption of the reverse color mapping operation for at least some sub-pictures.