3D Quasi-Recurrent Neural Network for Hyperspectral Image Denoising

TL;DR

This paper introduces a 3D quasi-recurrent neural network that effectively captures spatio-spectral correlations for hyperspectral image denoising, outperforming existing methods in accuracy and efficiency.

Contribution

It proposes an alternating directional 3D quasi-recurrent neural network that models both local and global correlations in hyperspectral images without additional computational cost.

Findings

Significant improvement over state-of-the-art methods in denoising accuracy.

Effective modeling of spatio-spectral dependencies.

Reduced computation time compared to existing approaches.

Abstract

In this paper, we propose an alternating directional 3D quasi-recurrent neural network for hyperspectral image (HSI) denoising, which can effectively embed the domain knowledge -- structural spatio-spectral correlation and global correlation along spectrum. Specifically, 3D convolution is utilized to extract structural spatio-spectral correlation in an HSI, while a quasi-recurrent pooling function is employed to capture the global correlation along spectrum. Moreover, alternating directional structure is introduced to eliminate the causal dependency with no additional computation cost. The proposed model is capable of modeling spatio-spectral dependency while preserving the flexibility towards HSIs with arbitrary number of bands. Extensive experiments on HSI denoising demonstrate significant improvement over state-of-the-arts under various noise settings, in terms of both restoration accuracy and computation time. Our code is available at https://github.com/Vandermode/QRNN3D.