P×Q Block Image Coding With Reduced-Dimension 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, generating orthogonal transform coefficients and quantizing them using an N×M array quantization matrix, with N<P and M<Q, to efficiently set some coefficients to zero, reducing calculation load and improving compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a large block size is used for orthogonal transform, then compression efficiency is improved, but the number of high-frequency coefficients that need to be processed increases, reducing coding efficiency
Solution Approach 1:
The patent extracts only the necessary low-frequency coefficients for processing while discarding or setting to zero the high-frequency coefficients. By using a quantization matrix with a specific structure where only a subset of coefficients are actively processed and others are implicitly zero, the system reduces the effective number of coefficients that need to be handled, thereby improving coding efficiency while maintaining compression benefits of large block sizes
2Measurement precision
If all orthogonal transform coefficients are processed, then image quality is maintained, but encoding time and computational load increase
Solution Approach 1:
The patent applies different processing quality to different parts of the frequency spectrum. Low-frequency coefficients receive full processing and quantization attention to maintain image quality, while high-frequency coefficients are handled differently (set to zero or processed with reduced precision). This local differentiation allows the system to maintain perceptual image quality while significantly reducing overall encoding time and computational load
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.