Machine vision-aware quality metrics for compressed image and video assessment

TL;DR

This paper develops new quality metrics tailored for compressed images and videos that better predict performance in machine vision tasks like detection and recognition, addressing the needs of automated systems.

Contribution

It introduces novel full-reference quality metrics specifically designed for machine vision applications, improving correlation with detection and recognition performance.

Findings

Proposed metrics outperform existing quality metrics in predicting machine vision task outcomes.

Metrics are tailored for objects, faces, and license plates detection and recognition.

Experimental results validate the effectiveness of the new metrics.

Abstract

A main goal in developing video-compression algorithms is to enhance human-perceived visual quality while maintaining file size. But modern video-analysis efforts such as detection and recognition, which are integral to video surveillance and autonomous vehicles, involve so much data that they necessitate machine-vision processing with minimal human intervention. In such cases, the video codec must be optimized for machine vision. This paper explores the effects of compression on detection and recognition algorithms (objects, faces, and license plates) and introduces novel full-reference image/video-quality metrics for each task, tailored to machine vision. Experimental results indicate our proposed metrics correlate better with the machine-vision results for the respective tasks than do existing image/video-quality metrics.