Fine-grained subjective visual quality assessment for high-fidelity compressed images

TL;DR

This paper introduces a new subjective assessment methodology for high-fidelity compressed images, utilizing boosting techniques and JND units to detect subtle quality differences, supported by a dataset and crowdsourced ratings.

Contribution

It presents a novel fine-grained quality assessment method for high-quality images, including a dataset, crowdsourced ratings, and a reconstruction approach in JND units.

Findings

High-precision quality scale in JND units achieved

Crowdsourced ratings effectively support fine-grained assessment

Boosting techniques improve artifact detection sensitivity

Abstract

Advances in image compression, storage, and display technologies have made high-quality images and videos widely accessible. At this level of quality, distinguishing between compressed and original content becomes difficult, highlighting the need for assessment methodologies that are sensitive to even the smallest visual quality differences. Conventional subjective visual quality assessments often use absolute category rating scales, ranging from ``excellent'' to ``bad''. While suitable for evaluating more pronounced distortions, these scales are inadequate for detecting subtle visual differences. The JPEG standardization project AIC is currently developing a subjective image quality assessment methodology for high-fidelity images. This paper presents the proposed assessment methods, a dataset of high-quality compressed images, and their corresponding crowdsourced visual quality ratings. It also outlines a data analysis approach that reconstructs quality scale values in just noticeable difference (JND) units. The assessment method uses boosting techniques on visual stimuli to help observers detect compression artifacts more clearly. This is followed by a rescaling process that adjusts the boosted quality values back to the original perceptual scale. This reconstruction yields a fine-grained, high-precision quality scale in JND units, providing more informative results for practical applications. The dataset and code to reproduce the results will be available at https://github.com/jpeg-aic/dataset-BTC-PTC-24.