Block Image Coding With Frequency-Specific Quantization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing High Efficiency Video Coding (HEVC) and Versatile Video Coding (VVC) methods face challenges in efficiently reducing encoding amounts while maintaining image quality, particularly in handling orthogonal transform coefficients of high-frequency components.
Innovation Solution
An image coding apparatus that encodes images in blocks of P×Q pixels, orthogonally transforms prediction residuals, and quantizes using N×M arrays of quantization matrices to generate quantized coefficients, with specific methods for zeroing out high-frequency components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If orthogonal transform coefficients of high-frequency components are forcibly set to 0 to reduce encoding amount, then compression efficiency is improved, but image quality deteriorates
Solution Approach 1:
The patent applies different quantization matrices to different regions of the frequency spectrum. Specifically, it uses a first quantization matrix for low-frequency components and a second quantization matrix for high-frequency components, allowing optimized handling of each frequency region independently to balance compression efficiency and image quality preservation
Solution Approach 2:
The patent dynamically adjusts quantization parameters by selecting between multiple quantization matrices based on the frequency components being processed. This parameter change strategy allows the system to adapt the quantization strength to the specific frequency band, reducing aggressive zeroing out for low frequencies while allowing more aggressive compression for high frequencies
2Productivity
If larger block size is used for orthogonal transform to improve coding efficiency, then compression performance is improved, but calculation amount increases
Solution Approach 1:
The patent divides the frequency spectrum into multiple bands (low-frequency and high-frequency components) and applies different quantization matrices to each segment. This segmentation allows the system to handle different frequency regions with appropriate complexity, reducing the overall calculation burden while maintaining coding efficiency
Solution Approach 2:
The patent applies partial quantization by selectively applying quantization matrices only to specific frequency components rather than uniformly to all coefficients. This partial action approach reduces the calculation amount by focusing computational resources only where necessary, while still achieving sufficient compression efficiency
Data Source
AI summary
An orthogonal transform unit orthogonally transforms prediction residuals in a block of a P×Q array of pixels, thereby generating an N×M (N is an integer satisfying N<P, and M is an integer satisfying M<Q) array of orthogonal transform coefficients. A quantization unit quantizes the N×M array of the orthogonal transform coefficients using at least a quantization matrix of an N×M array of elements, and thereby generates an N×M array of quantized coefficients.