Transfer CLIP for Generalizable Image Denoising

arXiv:2403.15132v1 Announce Type: new
Abstract: Image denoising is a fundamental task in computer vision. While prevailing deep learning-based supervised and self-supervised methods have excelled in eliminating in-distribution noise, their susceptibility to out-of-distribution (OOD) noise remains a significant challenge. The recent emergence of contrastive language-image pre-training (CLIP) model has showcased exceptional capabilities in open-world image recognition and segmentation. Yet, the potential for leveraging CLIP to enhance the robustness of low-level tasks remains largely unexplored. This paper uncovers that certain dense features …

abstract arxiv capabilities challenge clip computer computer vision cs.cv deep learning denoising distribution eess.iv emergence image language noise open-world pre-training training transfer type vision world