We hypothesize that optimal deep neural networks (DNN) class-conditional posterior probabilities live in a union of lowdimensional subspaces. In real test conditions, DNN posteriors encode uncertainties which can be regarded as a superposition of unstructured sparse noise over the optimal posteriors. We aim to investigate different ways to structure the DNN outputs by exploiting low-rank repres...

Low-Rank Representation (LRR) is arguably one of the most powerful paradigms for Multi-view spectral clustering, which elegantly encodes the multi-view local graph/manifold structures into an intrinsic low-rank self-expressive data similarity embedded in high-dimensional space, to yield a better graph partition than their single-view counterparts. In this paper we revisit it with a fundamentall...

The low-rank models have gained remarkable performance in the field of remote sensing image denoising. Nonetheless, existing low-rank-based methods view residues as noise and simply discard them. This causes denoised results to lose many important details, especially edges. In this paper, we propose a new denoising method named EPLRR-RSID, which focuses on edge preservation improve quality deta...

Tensor-based methods have been widely studied to attack inverse problems in hyperspectral imaging since a image (HSI) cube can be naturally represented as third-order tensor, which perfectly retain the spatial information image. In this article, we extend linear tensor method nonlinear and propose low-rank unmixing algorithm solve generalized bilinear model (GBM). Specifically, parts of GBM bot...