mlsanityData to 7 Oct 2026

An Unsupervised Information-Theoretic Perceptual Quality Metric

arXiv 2006.06752 ↗First posted 11 Jun 2020cs.CVcited by 28

Not on the map. Why?

Abstract (arXiv)

Tractable models of human perception have proved to be challenging to build. Hand-designed models such as MS-SSIM remain popular predictors of human image quality judgements due to their simplicity and speed. Recent modern deep learning approaches can perform better, but they rely on supervised data which can be costly to gather: large sets of class labels such as ImageNet, image quality ratings, or both. We combine recent advances in information-theoretic objective functions with a computational architecture informed by the physiology of the human visual system and unsupervised training on pairs of video frames, yielding our Perceptual Information Metric (PIM). We show that PIM is competitive with supervised metrics on the recent and challenging BAPPS image quality assessment dataset and outperforms them in predicting the ranking of image compression methods in CLIC 2020. We also perform qualitative experiments using the ImageNet-C dataset, and establish that PIM is robust with respect to architectural details.

Key facts. Cited by 28 papers in the map (data through 7 Oct 2026). First posted on arXiv 11 Jun 2020.

Builds on

No reference list yet.

Built on by

Cited by year

Citing papers per year: 2021: 4; 2022: 9; 2023: 5; 2024: 6; 2025: 4.

28 papers.

Sort:

By area:

  1. 2023Improving Statistical Fidelity for Neural Image Compression with Implicit Local Likelihood Modelscited by 90
  2. 2022An Introduction to Neural Data Compressioncited by 78
  3. 2021DICE: Diversity in Deep Ensembles via Conditional Redundancy Adversarial Estimationcited by 48
  4. 2022Shift-tolerant Perceptual Similarity Metriccited by 44
  5. 2022A Perceptual Quality Metric for Video Frame Interpolationcited by 15
  6. 2024Multiscale Sliced Wasserstein Distances as Perceptual Color Difference Measurescited by 14
  7. 2024Good, Cheap, and Fast: Overfitted Image Compression with Wasserstein Distortioncited by 13
  8. 2021On the relation between statistical learning and perceptual distancescited by 12
  9. 2022DeepDC: Deep Distance Correlation as a Perceptual Image Quality Evaluatorcited by 10
  10. 2022Approaching Peak Ground Truthcited by 8
  11. 2023Attacking Perceptual Similarity Metricscited by 8
  12. 2025ChatVis: Large Language Model Agent for Generating Scientific Visualizationscited by 7
16 more
  1. 2022Identifying and Mitigating Flaws of Deep Perceptual Similarity Metricscited by 4
  2. 2022Video Coding Using Learned Latent GAN Compressioncited by 4
  3. 2021Human Imperceptible Attacks and Applications to Improve Fairnesscited by 3
  4. 2024A Taxonomy of Miscompressions: Preparing Image Forensics for Neural Compressioncited by 3
  5. 2022Content-Diverse Comparisons improve IQAcited by 2
  6. 2023The Unreasonable Effectiveness of Linear Prediction as a Perceptual Metriccited by 2
  7. 2024Assessing invariance to affine transformations in image quality metricscited by 2
  8. 2021Curiously Effective Features for Image Quality Predictioncited by 1
  9. 2023Image Statistics Predict the Sensitivity of Perceptual Quality Metricscited by 1
  10. 2023CPIPS: Learning to Preserve Perceptual Distances in End-to-End Image Compressioncited by 1
  11. 2025Quality assessment of 3D human animation: Subjective and objective evaluationcited by 1
  12. 2025Non-Aligned Reference Image Quality Assessment for Novel View Synthesiscited by 1
  13. 2022Steerable Pyramid Transform Enables Robust Left Ventricle Quantificationcited by 0
  14. 2024Exploring Compressed Image Representation as a Perceptual Proxy: A Studycited by 0
  15. 2024Evaluating Perceptual Distance Models by Fitting Binomial Distributions to Two-Alternative Forced Choice Datacited by 0
  16. 2025Structured Uncertainty Similarity Score (SUSS): Learning a Probabilistic, Interpretable, Perceptual Metric Between Imagescited by 0