Context-Based Adaptive Variable Length Coding Position Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional Context-based Adaptive Variable Length Coding (CAVLC) for image and video compression inefficiently uses position coding schemes that do not account for the number of nonzero coefficients remaining, leading to suboptimal representation of quantized transform coefficients.
Innovation Solution
The proposed method employs a context-based adaptive variable length coding technique that accounts for both the number of zero-valued and nonzero-valued coefficients remaining, using mixed run-length position coding and integrated position coding methods to efficiently encode the positions of coefficients, ensuring that more likely sequences are represented by shorter codewords and less likely sequences by longer codewords.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional one-dimensional position coding scheme is used that depends only on the number of zeros remaining, then the coding process is simple, but the compression efficiency is suboptimal
Solution Approach 1:
The patent transitions from a one-dimensional position coding scheme (based only on number of zeros remaining) to a two-dimensional scheme that simultaneously considers both the number of zeros remaining and the number of nonzero coefficients remaining. This dimensional expansion enables more precise modeling of coefficient distribution patterns, improving compression efficiency by selecting VLC tables that better match the actual statistical characteristics of the transform coefficients.
Solution Approach 2:
The invention changes the parameters used for VLC table selection from a single parameter (number of zeros remaining) to two parameters (number of zeros remaining and number of nonzero coefficients remaining). This parameter change allows the coding system to adapt to different coefficient distribution scenarios more effectively, achieving better compression by matching the coding scheme to the actual data characteristics.
2Measurement precision
If the number of nonzero coefficients remaining is not used in the coding process, then the coding process is simpler, but the representation accuracy of coefficient positions is reduced
Solution Approach 1:
The patent incorporates feedback by using the number of nonzero coefficients remaining (along with zeros remaining) to dynamically select the most appropriate VLC table for encoding coefficient positions. This feedback mechanism allows the encoder to adapt to the actual distribution of coefficients in the current block, improving position representation accuracy by choosing coding tables that reflect the true statistical patterns rather than using a fixed or single-parameter-based scheme.
Data Source
AI summary
Particular embodiments include a method, an apparatus, and logic embodied in tangible computer-readable medium that when executed carries out a method of encoding an ordered sequence of quantized transform coefficients of a block of image data. One embodiment is a context adaptive variable length coding method that includes position coding the positions of zero-valued and non-zero valued coefficients by either a mixed method that encodes either the run length of zeroes preceding a non-zero coefficient or the run length of nonzero-valued coefficients preceding a zero-valued coefficients. Another includes position coding that uses a variable length code for two parameters respectively indicating the number of zero-valued coefficient positions and nonzero-valued coefficient positions still to be coded.