CASSI-Self-Supervised

This repository contains the codes for paper Self-supervised Neural Networks for Spectral Snapshot Compressive Imaging (ICCV (2021)) by Ziyi Meng, Zhenming Yu, KunXu, Xin Yuan. [pdf]

Overviewer

This repository uses a self-supervised neural networks to solve the reconstruction problem of snapshot compressive imaging (SCI), which uses a two-dimensional (2D) detector to capture a high-dimensional (usually 3D) data-cube in a compressed manner. This source code provides the reconstruction of 10 synthetic data originally used in TSA-Net paper. So far this version of code only includes the PnP-DIP for the synthetic data.

Results

Fig. 1 Reconstructed synthetic data (sRGB) by 8 algorithms. We show the reconstructed spectral curves on selected regions to compare the spectral accuracy of different algorithms.

Usage

Download the SMEM repository and model file

0. Requirements are Python 3 and Pytorch 1.6
1. Download this repository via git
2. Run main.py or main.ipynb to do reconstruction of one scene.

Citation

@article{meng2021self,
  title={Self-supervised Neural Networks for Spectral Snapshot Compressive Imaging},
  author={Meng, Ziyi and Yu, Zhenming and Xu, Kun and Yuan, Xin},
  journal={arXiv preprint arXiv:2108.12654},
  year={2021}
}

Contact

Ziyi Meng, Email: [email protected]

Xin Yuan, Westlake University, Email: [email protected]