mlsanityData to 7 Oct 2026

A Loss Function for Generative Neural Networks Based on Watson's Perceptual Model

arXiv 2006.15057 ↗First posted 26 Jun 2020cs.LG cs.CV eess.IVcited by 39

Not on the map. Why?

Abstract (arXiv)

To train Variational Autoencoders (VAEs) to generate realistic imagery requires a loss function that reflects human perception of image similarity. We propose such a loss function based on Watson's perceptual model, which computes a weighted distance in frequency space and accounts for luminance and contrast masking. We extend the model to color images, increase its robustness to translation by using the Fourier Transform, remove artifacts due to splitting the image into blocks, and make it differentiable. In experiments, VAEs trained with the new loss function generated realistic, high-quality image samples. Compared to using the Euclidean distance and the Structural Similarity Index, the images were less blurry; compared to deep neural network based losses, the new approach required less computational resources and generated images with less artifacts.

Key facts. Cited by 39 papers in the map (data through 7 Oct 2026). First posted on arXiv 26 Jun 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 3; 2022: 4; 2023: 10; 2024: 13; 2025: 7; 2026: 2.

39 papers.

Sort:

By area:

  1. 2023The Stable Signature: Rooting Watermarks in Latent Diffusion Modelscited by 289
  2. 2024WAVES: Benchmarking the Robustness of Image Watermarkscited by 89
  3. 2022Investigating Prompt Engineering in Diffusion Modelscited by 58
  4. 2024Video Seal: Open and Efficient Video Watermarkingcited by 45
  5. 2022Shift-tolerant Perceptual Similarity Metriccited by 44
  6. 2024Attack-Resilient Image Watermarking Using Stable Diffusioncited by 33
  7. 2023Explicitly Minimizing the Blur Error of Variational Autoencoderscited by 29
  8. 2024WMAdapter: Adding WaterMark Control to Latent Diffusion Modelscited by 29
  9. 2024WateRF: Robust Watermarks in Radiance Fields for Protection of Copyrightscited by 25
  10. 2024UnMarker: A Universal Attack on Defensive Image Watermarkingcited by 19
  11. 2024Boosting Latent Diffusion with Perceptual Objectivescited by 19
  12. 2021Evaluating the Interpretability of Generative Models by Interactive Reconstructioncited by 16
27 more
  1. 2021Clean Images are Hard to Reblur: Exploiting the Ill-Posed Inverse Task for Dynamic Scene Deblurringcited by 16
  2. 2024Stable Signature is Unstable: Removing Image Watermark from Diffusion Modelscited by 16
  3. 2022A Perceptual Quality Metric for Video Frame Interpolationcited by 15
  4. 2025Pixel Seal: Adversarial-only training for invisible image and video watermarkingcited by 14
  5. 2024LAVIB: A Large-scale Video Interpolation Benchmarkcited by 11
  6. 2021Semantic similarity metrics for learned image registrationcited by 9
  7. 2023Distributional Learning of Variational AutoEncoder: Application to Synthetic Data Generationcited by 9
  8. 2023FrePolad: Frequency-Rectified Point Latent Diffusion for Point Cloud Generationcited by 9
  9. 2023Attacking Perceptual Similarity Metricscited by 8
  10. 2024LVMark: Robust Watermark for Latent Video Diffusion Modelscited by 8
  11. 2022Simpler is better: spectral regularization and up-sampling techniques for variational autoencoderscited by 6
  12. 2024Towards Effective User Attribution for Latent Diffusion Models via Watermark-Informed Blendingcited by 4
  13. 2023AesFA: An Aesthetic Feature-Aware Arbitrary Neural Style Transfercited by 3
  14. 2025Video Signature: Implicit Watermarking for Video Diffusion Modelscited by 3
  15. 2023Enhancing image quality prediction with self-supervised visual maskingcited by 2
  16. 2024CSIM: A Copula-based similarity index sensitive to local changes for Image quality assessmentcited by 2
  17. 2024An Efficient Watermarking Method for Latent Diffusion Models via Low-Rank Adaptation and Dynamic Loss Weightingcited by 1
  18. 2025Pathology-Aware Adaptive Watermarking for Text-Driven Medical Image Synthesiscited by 1
  19. 2025VIDSTAMP: A Temporally-Aware Watermark for Ownership and Integrity in Video Diffusion Modelscited by 1
  20. 2025RePack then Refine: Efficient Diffusion Transformer with Vision Foundation Modelcited by 1
  21. 2023Explainable Recommender with Geometric Information Bottleneckcited by 0
  22. 2023Forest Parameter Prediction by Multiobjective Deep Learning of Regression Models Trained with Pseudo-Target Imputationcited by 0
  23. 2023Balanced Marginal and Joint Distributional Learning via Mixture Cramer-Wold Distancecited by 0
  24. 2025MultiNeRF: Multiple Watermark Embedding for Neural Radiance Fieldscited by 0
  25. 2025RAVQ-HoloNet: Rate-Adaptive Vector-Quantized Hologram Compressioncited by 0
  26. 2026Lossless Copyright Protection via Intrinsic Model Fingerprintingcited by 0
  27. 2026Rare-Aware Autoencoding: Reconstructing Spatially Imbalanced Datacited by 0