Article

Received 14.09.2025

Revised 02.12.2025

Accepted 29.12.2025

Retrieved from Vol. 28, No. 2, 2025

Pages 140 -150

Karnatov, S. (2025). Analysis of PSNR, SSIM, LPIPS metrics in the context of human perception of visual similarity. The National Transport University Bulletin: A Scientific and Technical Journal, 28(2), 140-150. https://doi.org/10.32703/2617-9040-2025-46-10

Analysis of PSNR, SSIM, LPIPS metrics in the context of human perception of visual similarity

Serhii Karnatov

This  paper  presents  a  comprehensive  comparative  analysis  of  three  well-known  image  quality assessment (IQA) metrics: PSNR, SSIM, and LPIPS. It explores their basic principles, mathematical foundations, advantages, and limitations, particularly as they relate to human visual perception. The evolution of IQA metrics from simple pixel-by-pixel comparisons (PSNR) to structural approaches (SSIM) and, more recently, to learned perceptual metrics (LPIPS) is discussed. A critical analysis of the effectiveness  of  each  metric  in  assessing  various  visual  distortions,  including  noise,  blur,  and compression artifacts, is presented. Inherent issues in human visual perception, such as the role of semantics, texture, color, and visual artifacts, are explored as fundamental causes of discrepancies between  objective  metric  estimates  and  subjective  human  judgments.  The  paper  highlights  the “unproven  effectiveness”  of  deep  features  in  LPIPS,  and  discusses  its  vulnerabilities,  such  as adversarial attacks and limitations in globalsemantic understanding. Finally, it outlines directions for future research aimed at developing more robust, interpretable, and perceptually consistent IQA metrics that can better account for the complexity of the human visual system and the evolving demands of modern image processing and generative artificial intelligence technologies

image quality assessment; PSNR; SSIM; LPIPS; human perception; visual distortions; generative models; objective metrics; subjective assessment
  1. Johnson, J., Alahi, A., & Fei-Fei, L. (2016, September). Perceptual losses for real-time style transfer and super-resolution. In European Conference on Computer Vision (pp. 694–711). Cham: Springer International Publishing. https://doi.org/10.1007/978-3-319-46475-6_43
  2. Wang, Z., Bovik, A. C., Sheikh, H. R., & Simoncelli, E. P. (2004). Image quality assessment: From error visibility to structural similarity. IEEE Transactions on Image Processing, 13(4), 600–612. https://doi.org/10.1109/TIP.2003.819861
  3. Breger, A., Biguri, A., Landman, M. S., Selby, I., Amberg, N., Brunner, E., … Schönlieb, C. B. (2025). A study of why we need to reassess full reference image quality assessment with medical images. Journal of Imaging Informatics in Medicine, 1–26. https://doi.org/10.1007/s10278-025-01462-1
  4. Zhang, R., Isola, P., Efros, A. A., Shechtman, E., & Wang, O. (2018). The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 586–595). https://arxiv.org/abs/1801.03924
  5. Arabboev, M., Begmatov, S., Rikhsivoev, M., Nosirov, K., & Saydiakbarov, S. (2024). A comprehensive review of image super-resolution metrics: Classical and AI-based approaches. Acta IMEKO, 13(1), 1–8. https://doi.org/10.21014/actaimeko.v13i1.1679
  6. Gertsiy, O. (2024). Research on graphic data formats for compact representation and comparison of images. Transport Systems and Technologies, (43), 173–187. https://doi.org/10.32703/2617-9059-2024-43-14
  7. Zhang, L., Zhang, L., Mou, X., & Zhang, D. (2011). FSIM: A feature similarity index for image quality assessment. IEEE Transactions on Image Processing, 20(8), 2378–2386. https://doi.org/10.1109/TIP.2011.2109730
  8. Shrestha, B. (2005). Evaluation of JPEG2000 for lossless medical image compression. Mississippi State University Libraries.
  9. Russ, J. C. (2006). The Image Processing Handbook. CRC Press.
  10. Gonzalez, R. C., & Woods, R. E. (2017). Digital Image Processing (4th ed.). Pearson, New York.
  11. Singh, G. K., Agarwal, A., & Reddy, N. V. (2023). Comparison of PSNR, SSIM, and LPIPS in medical imaging. In 2023 IEEE 12th International Conference on Computing Communication and Networking Technologies (ICCCNT) (pp. 1–6).
  12. Wang, Z., Simoncelli, E. P., & Bovik, A. C. (2003, November). Multiscale structural similarity for image quality assessment. In The Thirty-Seventh Asilomar Conference on Signals, Systems & Computers (Vol. 2, pp. 1398–1402). IEEE. https://doi.org/10.1109/ACSSC.2003.1292216
  13. Kuzovkin, I., Vicente, R., Petton, M., Lachaux, J.-P., Baciu, M., Kahane, P., … Aru, J. (2018). Activations of deep convolutional neural networks are aligned with gamma band activity of human visual cortex. Communications Biology, 1(1), 107. https://doi.org/10.1038/s42003-018-0110-y
  14. Zhang, K., Liang, J., Van Gool, L., & Timofte, R. (2021). Designing a practical degradation model for deep blind image super-resolution. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 4791–4800). https://doi.org/10.48550/arXiv.2103.14006
  15. Zhai, G., & Min, X. (2020). Perceptual image quality assessment: A survey. Science China Information Sciences, 63(11), 211301. https://doi.org/10.1007/s11432-019-2757-1
  16. Gu, S., Bao, J., Chen, D., & Wen, F. (2020, August). GIQA: Generated image quality assessment. In European Conference on Computer Vision (pp. 369–385). Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030-58621-8_22
  17. Shoshan, A., Gandelsman, Y., Bagon, S., & Dekel, T. (2024). R-LPIPS: An adversarially robust perceptual similarity metric. Scientific Reports. https://doi.org/10.48550/arXiv.2307.15157
  18. Koffka, K. (1935). Principles of Gestalt Psychology. Harcourt, Brace & Co.
  19. Marr, D. (1982). Vision: A computational investigation into the human representation and processing of visual information. W. H. Freeman and Company.
  20. International Telecommunication Union. (2019). Methodology for the subjective assessment of the quality of television pictures (Recommendation BT.500-14).

https://doi.org/10.32703/2617-9040-2025-46-10