Residual Dense Autoencoder Network for Nonlinear Hyperspectral Unmixing

Abstract

Hyperspectral unmixing is a popular research topic in hyperspectral processing, aiming at obtaining the ground features contained in the mixed pixels and their proportion. Recently, nonlinear mixing models have received particular attention in hyperspectral decomposition since the linear mixing model cannot suitably apply in the situation that exists in multiple scattering. In this study, we constructed a residual dense autoencoder network (RDAE) for nonlinear hyperspectral unmixing in multiple scattering scenarios. First, an encoder was built based on the residual dense network (RDN) and attention layer. The RDN is employed to characterize multi-scale representations, which are further transformed with the attention layer to estimate the abundance maps. Second, we designed a decoder based on the unfolding of a generalized bilinear model to extract endmembers and estimate their second-order scattering interactions. Comparative experiments between the RDAE and six other state-of-the-art methods under synthetic and real hyperspectral datasets demonstrate that the proposed method achieved a better performance in terms of endmember extraction and abundance estimation.

IVIU-Net: Implicit Variable Iterative Unrolling Network for Hyperspectral Sparse Unmixing

“…It was trained by a loss function including a quadratic term based on the Hapke model, the reconstruction error of the reflectances, and a minimum volume total variation (TV) term. A residual dense autoencoder network is constructed in [38] for nonlinear decomposition in multiple scattering scenes. And a new type of deep automatic encoder network is designed based on the generalized bilinear model in [39].…”

Enhanced Lung Nodule Segmentation using Dung Beetle Optimization based LNS-DualMAGNet Model

“…Two main parts make up the DualMAF module: channel attention methods and atrous convolution with various expansion rates. Rich feature information can be captured by the atrous convolution method, which can increase the convolution kernel's receptive field without changing the computational parameters [28]. Moreover, channel attention is added to reinforce the interaction between several sequences in order to enhance the feature recognition ability.…”