Image Compression Using Adaptive Dead Zone Quantization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image compression methods often control compression quality through bit rate, limiting local regulation and being unsuitable for applications like satellite image transmission where quality varies across image zones without constrained bit rates.
Innovation Solution
An image compression method using a distortion model independent of bit rate, involving decorrelation, block quantization with a uniform scalar quantizer having a dead zone, and adaptive quantization step determination to achieve target quality in each image block.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If image compression quality is controlled via bit rate, then global compression control is achieved, but local quality regulation is lost
Solution Approach 1:
The image is divided into multiple blocks, and each block is independently quantized with its own dead zone parameter T. This segmentation enables local quality regulation while maintaining overall compression control, directly resolving the contradiction between global control and local adaptability.
Solution Approach 2:
Different blocks of the image can have different dead zone parameters T, allowing each region to be compressed with appropriate quality levels according to its importance. This local quality approach enables selective compression where critical regions maintain higher quality while less important regions use higher compression.
2Manufacturing precision
If complex iterative processes are used for compression, then compression quality can be optimized, but device complexity increases
Solution Approach 1:
The distortion model is pre-established with closed-form expressions for quantization distortion D(T) and bit rate R(T). This preliminary preparation of mathematical models allows direct computation of optimal parameters without iterative processes, reducing implementation complexity while maintaining optimization capability.
Solution Approach 2:
The invention transforms the complex iterative optimization problem into a simple parameter selection problem by deriving closed-form relationships between the dead zone parameter T and both distortion and bit rate. This parameter transformation approach enables quality optimization through direct calculation rather than complex iteration.
3Productivity
If distortion modeling depends on bit rate, then compression efficiency is improved, but quality control without bit rate constraints becomes difficult
Solution Approach 1:
Instead of modeling distortion as a function of bit rate, the invention inverts the approach by modeling both distortion D(T) and bit rate R(T) as functions of the dead zone parameter T. This inversion enables quality control through distortion minimization independent of bit rate constraints, while still maintaining compression efficiency.
Data Source
Figure 1
Figure 2a~3
AI summary
Image compression method comprising the following steps applied to at least one image (I): • Decouple (101) the image (I) by applying a mathematical transform to it so as to obtain a set of coefficients, • Decompose (102) the image (I) into blocks and, for each block of the image (I), • Quantize (105) said coefficients using a uniform dead-zone scalar quantizer having a half-dead zone of size T and a quantization step Δ, • Encode (106) the quantized coefficients, • The size T of the half-dead zone of the scalar quantizer being determined (103) by minimizing the gap between an estimated quantization distortion D(T), depending at least on said size T, and a target quantization distortion Dc.