Tensorflow codes for paper: Shipeng Zhang, Lizhi Wang, Lei Zhang, and Hua Huang, Learning Tensor Low-Rank Prior for Hyperspectral Image Reconstruction, IEEE CVPR, 2021. [pdf]
Snapshot hyperspectral imaging has been developed to capture the spectral information of dynamic scenes. In this paper, we propose a deep neural network by learning the tensor low-rank prior of hyperspectral images (HSI) in the feature domain to promote the reconstruction quality. Our method is inspired by the canonical-polyadic (CP) decomposition theory, where a low-rank tensor can be expressed as a weight summation of several rank-1 component tensors. Specifically, we first learn the tensor low-rank prior of the image features with two steps: (a) we generate rank-1 tensors with discriminative components to collect the contextual information from both spatial and channel dimensions of the image features; (b) we aggregate those rank-1 tensors into a low-rank tensor as a 3D attention map to exploit the global correlation and refine the image features. Then, we integrate the learned tensor low-rank prior into an iterative optimization algorithm to obtain an end-to-end HSI reconstruction. Experiments on both synthetic and real data demonstrate the superiority of our method.
We provide data of two datasets (Harvard and ICVL) for training and testing. It can be download from Data.
The resolution of testing data provided here is
Python 2.7.18
CUDA 9.0
Tensorflow 1.11.0
- Download this repository via git or download the zip file manually.
git clone https://github.com/zspCoder/DTLP
-
Download the data from Data
-
Run the file Train.py to train the model.
-
Run the file Test.py to test the model.
@inproceedings{zhang2021learning,
title={Learning Tensor Low-Rank Prior for Hyperspectral Image Reconstruction},
author={Zhang, Shipeng and Wang, Lizhi and Zhang, Lei and Huang, Hua},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={12006--12015},
year={2021}
}