LDR Sample Mapping for Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently encoding and decoding images and videos using Low Dynamic Range (LDR) samples, particularly in reducing redundancy and optimizing bitstream signaling.
Innovation Solution
The implementation of sample mapping functions, including lookup tables, algorithmic operations, and deep learning models, to convert samples from an original bit depth to an LDR bit depth during encoding and decoding, with signaling of these mapping functions in the bitstream.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If LDR sample mapping is applied to convert samples from original bit depth to LDR bit depth, then coding efficiency is improved and bitstream overhead is reduced, but device complexity increases due to the need for mapping functions and lookup tables
Solution Approach 1:
The patent applies LDR sample mapping during the encoding process before compression, converting samples from original bit depth to LDR bit depth using predefined or signaled mapping functions. This preliminary conversion simplifies subsequent coding operations and reduces bitstream overhead, while the decoder applies the inverse mapping to reconstruct samples at the original bit depth.
Solution Approach 2:
The patent changes the bit depth parameter of video samples by applying LDR mapping functions that transform sample values from one bit depth range to another. This parameter transformation enables more efficient compression at reduced bit depths while allowing reconstruction at higher bit depths, effectively managing the trade-off between coding efficiency and quality.
2Productivity
If bit depth of video samples is reduced to increase coding efficiency, then loss of information increases, but coding efficiency is improved
Solution Approach 1:
The patent applies LDR mapping functions that non-linearly transform sample values, concentrating information in the most significant bits while reducing the dynamic range. This allows efficient encoding at lower bit depths by optimizing the distribution of sample values, thereby reducing information loss compared to simple truncation or uniform quantization.
Data Source
AI summary
This disclosure relates generally to video coding and particularly to methods and systems for sample mapping in image and video coding using Low Dynamic Range (LDR) samples. For example, the disclosure describes implementations of mapping functions for conversion of samples from an original bit depth to an LDR bit depth during an encoding and decoding process of videos or images and the signaling of such mapping functions in a bitstream.