Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

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

Michela Testolina, Mohsen Jenadeleh, Shima Mohammadi, Shaolin Su, Joao Ascenso, Touradj Ebrahimi, Jon Sneyers, Dietmar Saupe

TL;DR

The proposed assessment methods are presented, a dataset of high-quality compressed images, and their corresponding crowdsourced visual quality ratings are presented, and a data analysis approach that reconstructs quality scale values in just noticeable difference (JND) units is outlined.

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.

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

TL;DR

The proposed assessment methods are presented, a dataset of high-quality compressed images, and their corresponding crowdsourced visual quality ratings are presented, and a data analysis approach that reconstructs quality scale values in just noticeable difference (JND) units is outlined.

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.

Paper Structure

This paper contains 14 sections, 3 figures.

Table of Contents

  1. Introduction
  2. Related work
  3. Proposed JPEG AIC-3 test methodologies
  4. Experimental setup
  5. Data analysis
  6. Experimental results and discussion
  7. Conclusions and future work
  8. Acknowledgments
  9. References

Figures (3)

  • Figure 1: Sketch of the interface of the BTC and PTC experiment.
  • Figure 2: Crops of the reference images adopted in the experiment.
  • Figure 5: Psychometric functions and JND threshold determination with and without the artifact boosting. The ratio of correct responses to same-codec triplet questions comparing distortion levels 0 and 10 averaged over all sources and codecs. The JND threshold, according to the AIC-2 flicker test, is between the two dotted lines. The dark-shaded region, therefore, is in the visually lossless region, and the lightly shaded region is partly visually lossless.