2020/06/26 by Steffen Czolbe, Czolbe, Steffen, Oswin Krause +6 · 14 citations
Computer Science · Engineering · Mathematics · #Advanced Image Processing Techniques #Artificial intelligence #Artificial neural network #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Computer vision #FOS: Computer and information sciences #FOS: Electrical engineering #Generative Adversarial Networks and Image Synthesis #Image (mathematics) #Image and Signal Denoising Methods #Image and Video Processing (eess.IV) #Luminance #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Pattern recognition (psychology) #Robustness (evolution) #Similarity (geometry) #cs.CV #cs.LG #eess.IV #electronic engineering #information engineering #stat.ML
paper · pdf · doi:10.48550/arxiv.2006.15057
published in arXiv (Cornell University) (Cornell University) · Published at the 34th Conference on Neural Information Processing Systems (NeurIPS 2020)
openalex publication_date 2020/06/26 · arxiv created 2021/01/06 · arxiv updated 2021/01/07 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05
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.