Progressive Lossless Video Coding via Layered H.264 Compatibility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies, such as JPEG2000, JPEG-LS, and H.264's Fidelity Range Extension, face inefficiencies in lossless coding due to intra-frame coding, limited progressive transmission, and inefficiencies in orthogonal transformation and quantization, which hinder high compression efficiency and compatibility with the H.264 standard.
Innovation Solution
A progressive lossless video coding method that involves inputting residual signals from image blocks, applying orthogonal transformation, quantization, and validity judgment to enumerate and code grid points within a predetermined existential space, allowing for efficient lossless decoding while maintaining compatibility with the H.264 standard and minimizing additional coding amount.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If intra-frame coding is used (JPEG2000, JPEG-LS), then lossless coding is achieved, but inter-frame correlation cannot be utilized resulting in lower compression efficiency
Solution Approach 1:
The video coding process is segmented into two distinct parts: a base layer using conventional lossy H.264 coding, and an enhancement layer using lossless coding on prediction residuals. This segmentation allows each layer to optimize for its specific function while combining benefits of both approaches.
Solution Approach 2:
The invention transitions from traditional single-layer coding to a two-dimensional layered structure (lossy base layer + lossless enhancement layer). By adding this dimensional layering, the system can simultaneously achieve lossless accuracy and high compression efficiency through inter-frame correlation in the enhancement layer.
2Manufacturing precision
If conventional H.264 Fidelity Range Extension is used, then lossless coding is attempted, but it cannot transmit progressively and has limited efficiency
Solution Approach 1:
The coding system is made dynamic and adjustable through layered structure. The enhancement layer can be independently controlled and transmitted progressively, allowing flexible adaptation to different bandwidth and decoding requirements while maintaining lossless accuracy when fully received.
3Productivity
If DCT transformation is applied followed by integer transformation (MPEG-4 FGS), then coding efficiency is enhanced, but lossless coding becomes impossible due to information loss
Solution Approach 1:
Lossless coding is applied preliminarily to the prediction residuals before they are combined with the lossy base layer. By performing lossless coding on the difference signal first, the original information is preserved in the enhancement layer, enabling perfect reconstruction when both layers are decoded.
Solution Approach 2:
The prediction residual signal serves as an intermediary between the lossy base layer and the lossless enhancement layer. This intermediary contains the difference information that, when coded losslessly and added to the base layer output, enables perfect reconstruction of the original signal.
4Productivity
If inter-frame prediction is used to enhance coding efficiency, then compression is improved, but compatibility with H.264 standard and scalability are compromised
Solution Approach 1:
The enhancement layer is designed with universal applicability that works with any H.264 compliant base layer. The lossless coding mechanism can be applied to both intra-frame and inter-frame predictions, making the system universally compatible while maintaining scalability and standard adherence.
Data Source
AI summary
Highly efficient lossless decoding is realized under the condition that codes transmitted as a base part are compatible with the H.264 standard. An orthogonal transformation section (12) performs orthogonal transformation of residual signals (Rorig), acquires transform coefficients (Xorig), and a quantization section (13) quantizes the transform coefficients. An existential space determination section (14) obtains information on upper limits and lower limits of the respective coefficients (an existential space of transform coefficients) from quantization information. A bundled coefficients coding section (16) decides whether respective grid points in the existential space of the transform coefficients have validity as results of orthogonal transformation of the residual signals, and enumerates valid grid points, assigns serial numbers (index) in the order of enumeration, and encodes using a serial number coding section (166) the serial numbers of grid points which match the transform coefficients (Xorig) of the residual signal.