Invertible Denoising Network: A Light Solution for Real Noise Removal
This work addresses image denoising for applications requiring lightweight models, though it appears incremental as it builds on invertible networks for a known bottleneck.
The authors tackled the problem of real noise removal in images by proposing an invertible denoising network (InvDN), which transforms noisy inputs into clean images and latent noise representations, achieving state-of-the-art results on the SIDD dataset with significantly fewer parameters (4.2% of DANet) and less run time.
Invertible networks have various benefits for image denoising since they are lightweight, information-lossless, and memory-saving during back-propagation. However, applying invertible models to remove noise is challenging because the input is noisy, and the reversed output is clean, following two different distributions. We propose an invertible denoising network, InvDN, to address this challenge. InvDN transforms the noisy input into a low-resolution clean image and a latent representation containing noise. To discard noise and restore the clean image, InvDN replaces the noisy latent representation with another one sampled from a prior distribution during reversion. The denoising performance of InvDN is better than all the existing competitive models, achieving a new state-of-the-art result for the SIDD dataset while enjoying less run time. Moreover, the size of InvDN is far smaller, only having 4.2% of the number of parameters compared to the most recently proposed DANet. Further, via manipulating the noisy latent representation, InvDN is also able to generate noise more similar to the original one. Our code is available at: https://github.com/Yang-Liu1082/InvDN.git.