Image Compression With Adaptive Quantization for Controlled Quality Loss

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image compression methods apply uniform quantization strategies across the entire image, leading to suboptimal quality loss and artifacts, as they fail to account for the varying content and characteristics of different areas within an image.

Innovation Solution

Implement a sequence of quality control metrics to iteratively process transform coefficients based on a defined error level, allowing for non-linear quality loss control by applying different quantization decisions to various image areas, thus minimizing bit cost and quality loss.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If uniform quantization strategy is applied across the entire image, then the compression process is simple and fast, but the image quality loss is suboptimal and artifacts are increased

Engineering Contradiction:
Improvecompression speedVSAvoidimage quality
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The image is divided into multiple regions based on content characteristics (e.g., smooth areas, textured areas, edge areas). Each region is then processed with appropriate quantization parameters tailored to its specific characteristics, allowing for optimized quality preservation in critical regions while maintaining compression efficiency in less critical regions.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different quantization strategies and parameters are applied to different regions of the image based on local content characteristics. Important regions (such as faces, text, or edges) receive higher quality treatment with finer quantization, while less important regions use coarser quantization, achieving optimal overall quality for the given bit rate.

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If different quantization strategies are applied to different image areas, then the image quality is optimized, but the compression process becomes more complex

Engineering Contradiction:
Improveimage qualityVSAvoidcompression process complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The image is pre-processed to identify and classify different regions based on content characteristics before the actual compression process. This preliminary segmentation and classification enables the subsequent compression stage to efficiently apply appropriate quantization parameters without excessive complexity during the main compression operation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The compression algorithm automatically analyzes the image content and self-adjusts quantization parameters for different regions without requiring manual intervention or complex external control systems. The algorithm itself performs the region classification and parameter selection, simplifying the overall system architecture.

Inventive Principle:
Principle #25Self-service

3Loss of time

If higher compression is applied to reduce data size, then the transmission and rendering time is reduced, but the perceptual quality degradation increases

Engineering Contradiction:
Improvetransmission and rendering timeVSAvoidperceptual quality
Core Design Contradiction:
Loss of timeVSLoss of information

Solution Approach 1:

The quantization parameters are dynamically adjusted based on the desired bit rate and image content characteristics. By changing parameters such as quantization step size, precision, and rounding modes adaptively, the system achieves optimal compression ratios while minimizing perceptual quality loss for the given bandwidth constraints.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3732881B1Image compression using controlled quality loss
Publication Date: 2025.07.30 GOOGLE LLC
  • EP3732881B1 patent drawingFigure 1
  • EP3732881B1 patent drawingFigure 2
  • EP3732881B1 patent drawingFigure 3

AI summary

The loss of image quality during compression is controlled using a sequence of quality control metrics. The sequence of quality control metrics is selected for quantizing transform coefficients within an area of the image based on an error level definition. Candidate bit costs are then determined by quantizing the transform coefficients according to the error level definition or a modified error level and the sequence of quality control metrics. Where the candidate bit cost resulting from using the modified error level is lower than the candidate bit cost resulting from using the error level definition, the transform coefficients are quantized according to the modified error level and the sequence of quality control metrics. Otherwise, the transform coefficients are quantized based on the error level definition and according to the sequence of quality control metrics.