Image Coding With Selective High-Frequency Coefficient Zeroing
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
1Device complexity
If orthogonal transform coefficients of high-frequency components are forcibly set to 0 to reduce encoding amount, then encoding complexity is reduced, but image quality may deteriorate
Solution Approach 1:
The patent applies local quality by differentiating the handling of orthogonal transform coefficients based on their frequency characteristics. High-frequency coefficients are forcibly set to 0 to reduce encoding complexity, while low-frequency coefficients are preserved to maintain image quality. This selective processing allows different parts of the coefficient data to be treated differently according to their importance for image quality.
Solution Approach 2:
The patent changes the parameter of coefficient values by forcibly setting high-frequency orthogonal transform coefficients to 0. This parameter modification reduces the amount of data that needs to be encoded, thereby reducing encoding complexity and bit rate, while the impact on image quality is minimized through the selective zeroing of only high-frequency components.
2Productivity
If larger block size is used for orthogonal transform to improve coding efficiency, then coding efficiency is improved, but calculation amount increases
Solution Approach 1:
The patent extracts and removes high-frequency orthogonal transform coefficients by forcibly setting them to 0. This extraction reduces the amount of data that needs to be processed and encoded, thereby reducing the calculation amount while maintaining the benefits of using larger block sizes for improved coding efficiency.
Solution Approach 2:
The patent applies partial action by selectively processing only the high-frequency coefficients for zeroing out, rather than processing all coefficients uniformly. This partial processing approach reduces the calculation amount significantly while still achieving the coding efficiency benefits of larger block transforms.
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.