2018/05/08 by Yu Liu, Yü Liu, Liu, Yu +2 · 4 citations
Computer Science · Engineering · Mathematics · Neuroscience · Psychology · #Advanced Image Fusion Techniques #Advanced Image Processing Techniques #Artificial intelligence #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Computer vision #FOS: Computer and information sciences #Geometry #Image (mathematics) #Image Enhancement Techniques #Mathematics #Net (polyhedron) #Neuroscience #Perception #Psychology #cs.CV
paper · pdf · doi:10.48550/arxiv.1805.03146
published in arXiv (Cornell University) (Cornell University) · 8 pages, 4 figures; project page: https://github.com/guanlongzhao/single-image-dehazing
arxiv created 2018/05/08 · openalex publication_date 2018/05/08 · arxiv updated 2018/05/09 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
In this work, we investigate the possibility of replacing the ℓ2 loss with perceptually derived loss functions (SSIM, MS-SSIM, etc.) in training an end-to-end dehazing neural network. Objective experimental results suggest that by merely changing the loss function we can obtain significantly higher PSNR and SSIM scores on the SOTS set in the RESIDE dataset, compared with a state-of-the-art end-to-end dehazing neural network (AOD-Net) that uses the ℓ2 loss. The best PSNR we obtained was 23.50 (4.2% relative improvement), and the best SSIM we obtained was 0.8747 (2.3% relative improvement.)