Projects per year
Abstract
This paper investigates the problem of recovering hyperspectral (HS) images from single RGB images. To tackle such a severely ill-posed problem, we propose a physically-interpretable, compact, efficient and end-to-end learning-based framework, namely AGD-Net. Precisely, by taking advantage of the imaging process, we first formulate the problem explicitly based on the classic gradient descent algorithm. Then, we design a lightweight neural network with a multi-stage architecture to mimic the formed amended gradient descent process, in which efficient convolution and novel spectral zero-mean normalization are proposed to effectively extract spatial-spectral features for regressing an initialization, a basic gradient, and an incremental gradient. Besides, based on the approximate low-rank property of HS images, we propose a novel rank loss to promote the similarity between the global structures of reconstructed and ground-truth HS images, which is optimized with our singular value weighting strategy during training. Moreover, AGD-Net, a single network after one-time training, is flexible to handle the reconstruction with various spectral response functions. Extensive experiments over three commonly-used benchmark datasets demonstrate that AGD-Net can improve the reconstruction quality more than 1.0 dB on average while saving 67x parameters and 32x FLOPs, compared with state-of-the-art methods. The code will be publicly available at https://github.com/zbzhzhy/GD-Net.
Original language | English |
---|---|
Pages (from-to) | 1176-1188 |
Journal | IEEE Transactions on Computational Imaging |
Volume | 7 |
Online published | 2 Nov 2021 |
DOIs | |
Publication status | Published - 2021 |
Research Keywords
- Hyperspectral imagery
- spectral reconstruction
- deep learning
- gradient descent
- rank loss
Fingerprint
Dive into the research topics of 'Deep Amended Gradient Descent for Efficient Spectral Reconstruction from Single RGB Images'. Together they form a unique fingerprint.Projects
- 1 Finished
-
GRF: Learning Based Hyperspectral Image Reconstruction and Discriminative Representation
HOU, J. (Principal Investigator / Project Coordinator)
1/01/20 → 22/12/23
Project: Research
Student theses
-
Learning-based High-quality Hyperspectral Images Reconstruction
ZHU, Z. (Author), HOU, J. (Supervisor), 12 Jul 2023Student thesis: Doctoral Thesis