Latest revision as of 23:13, 1 February 2021

Abstract

Recovering a high dynamic range (HDR) image from a single low dynamic range (LDR) input image is challenging due to missing details in under-/over-exposed regions caused by quantization and saturation of camera sensors. In contrast to existing learning-based methods, our core idea is to incorporate the domain knowledge of the LDR image formation pipeline into our model. We model the HDRto-LDR image formation pipeline as the (1) dynamic range clipping, (2) non-linear mapping from a camera response function, and (3) quantization. We then propose to learn three specialized CNNs to reverse these steps. By decomposing the problem into specific sub-tasks, we impose effective physical constraints to facilitate the training of individual sub-networks. Finally, we jointly fine-tune the entire model end-to-end to reduce error accumulation. With extensive quantitative and qualitative experiments on diverse image datasets, we demonstrate that the proposed method performs favorably against state-of-the-art single-image HDR reconstruction algorithms.

Comment: CVPR 2020. Project page: https://www.cmlab.csie.ntu.edu.tw/~yulunliu/SingleHDR Code: https://github.com/alex04072000/SingleHDR

Original document

The different versions of the original document can be found in:

• http://arxiv.org/abs/2004.01179

• http://arxiv.org/pdf/2004.01179

• http://xplorestaging.ieee.org/ielx7/9142308/9156271/09157653.pdf?arnumber=9157653,

http://dx.doi.org/10.1109/cvpr42600.2020.00172

• https://dblp.uni-trier.de/db/conf/cvpr/cvpr2020.html#LiuLCK0CH20,

https://arxiv.org/abs/2004.01179,

https://openaccess.thecvf.com/content_CVPR_2020/html/Liu_Single-Image_HDR_Reconstruction_by_Learning_to_Reverse_the_Camera_Pipeline_CVPR_2020_paper.html,

https://openaccess.thecvf.com/content_CVPR_2020/papers/Liu_Single-Image_HDR_Reconstruction_by_Learning_to_Reverse_the_Camera_Pipeline_CVPR_2020_paper.pdf,

https://arxiv.org/pdf/2004.01179,

https://academic.microsoft.com/#/detail/3034628923