Image Encoding Region Compression for Machine Vision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image compression technologies face challenges in achieving high compression rates while maintaining the performance of machine vision tasks such as object detection, image division, and object tracking, which are essential in applications like smart cities and autonomous driving.
Innovation Solution
An image encoding/decoding method and device that determines optimal compression levels for regions of interest within an image by adjusting resolution and quantization levels based on properties like size, pixel values, and proximity, using a compression rate control algorithm and probability distribution models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If image compression rate is increased, then storage and transmission efficiency is improved, but machine vision task performance deteriorates
Solution Approach 1:
The image is divided into multiple regions of interest (ROIs) based on machine vision task requirements. Different compression levels are applied to different ROIs, allowing critical regions to maintain higher quality for accurate machine vision processing while non-critical regions use higher compression to improve overall compression rate.
Solution Approach 2:
Different quality levels (compression rates) are applied locally to different regions of the image based on their importance for machine vision tasks. Critical regions maintain high quality with low compression, while less important regions use high compression, optimizing the balance between compression rate and task performance.
2Ease of operation
If uniform compression is applied to the entire image, then processing simplicity is improved, but machine vision performance in critical regions deteriorates
Solution Approach 1:
The image is segmented into multiple regions of interest with different compression requirements. This segmentation allows the system to apply different compression strategies to different regions, improving detection accuracy in critical areas while maintaining overall processing efficiency through automated region classification.
Solution Approach 2:
Different compression quality levels are applied to different regions based on their importance for machine vision tasks. Critical regions receive higher quality (lower compression) to maintain detection precision, while non-critical regions use higher compression, optimizing the trade-off between processing simplicity and detection accuracy.
3Loss of energy
If high compression is applied to all regions, then data transmission efficiency is improved, but information quality for vision tasks deteriorates
Solution Approach 1:
The image data is segmented into critical and non-critical regions. Critical regions are transmitted with lower compression to preserve information quality for machine vision tasks, while non-critical regions use higher compression to improve overall transmission efficiency, reducing total data transmission requirements.
Solution Approach 2:
Different compression quality levels are applied locally to different regions based on their information importance. Regions critical for machine vision tasks maintain high information quality with lower compression, while less important regions use higher compression to improve overall transmission efficiency and reduce data loss in critical areas.
Data Source
AI summary
An image encoding/decoding method, device and recording medium based sed on multiple compression levels disclosure may include extracting a region of interest for machine vision from an input image, determining a compression level of the region of interest, and encoding the compression level of the region of interest.