Disentangling Noise from Images: A Flow-Based Image Denoising Neural Network.

The prevalent convolutional neural network (CNN)-based image denoising methods extract features of images to restore the clean ground truth, achieving high denoising accuracy. However, these methods may ignore the underlying distribution of clean images, inducing distortions or artifacts in denoising results. This paper proposes a new perspective to treat image denoising as a distribution learning and disentangling task. Since the noisy image distribution can be viewed as a joint distribution of clean images and noise, the denoised images can be obtained via manipulating the latent representations to the clean counterpart. This paper also provides a distribution-learning-based denoising framework. Following this framework, we present an invertible denoising network, FDN, without any assumptions on either clean or noise distributions, as well as a distribution disentanglement method. FDN learns the distribution of noisy images, which is different from the previous CNN-based discriminative mapping. Experimental results demonstrate FDN's capacity to remove synthetic additive white Gaussian noise (AWGN) on both category-specific and remote sensing images. Furthermore, the performance of FDN surpasses that of previously published methods in real image denoising with fewer parameters and faster speed.

An HRD/DER-independent ER quality control mechanism involves Rsp5p-dependent ubiquitination and ER-Golgi transport.

We have identified a new pathway of ER-associated degradation in Saccharomyces cerevisiae that functions separately from the HRD/DER pathway comprised of Hrd1p, Hrd3p, Der1p, and Ubc7p. This pathway, termed Hrd1p independent-proteolysis (HIP), is capable of recognizing and degrading both lumenal (CPY* and PrA*), and integral membrane proteins (Sec61-2p) that misfold in the ER. CPY* overexpression likely saturates the HRD/DER pathway and activates the HIP pathway, so the slowed degradation kinetics of CPY* in a hrd1 Delta strain is restored to a wild-type rate when CPY* is overexpressed. Substrates of HIP require vesicular trafficking between the ER and Golgi apparatus before degradation by the ubiquitin-proteasome system. Ubiquitination of HIP substrates does not involve the HRD/DER pathway ubiquitin ligase Hrd1p, but instead uses another ubiquitin ligase, Rsp5p. HIP is regulated by the unfolded protein response as Ire1p is necessary for the degradation of CPY* when overexpressed, but not when CPY* is expressed at normal levels. Both the HIP and HRD/DER pathways contribute to the degradation of CPY*, and only by eliminating both is CPY* degradation completely blocked.