We present a snapshot multispectral imaging technique with a computational camera and a corresponding image restoration algorithm. The main characteristics are: (1) transferring spectral information to the spatial domain by engineering user-defined PSFs; (2) measuring spectral images by computationally inverting the image formation. The design of our computational camera is based on a phase-coded aperture technique to generate spatial and spectral variant PSFs. The corresponding algorithm is designed by adapting single-channel and cross-channel priors. We show experimentally the viability of our technique: it reconstructs high-resolution multispectral images from a snapshot. We further validate that the role of PSF design is critical.
This repository contains the reconstruction MATLAB code from the encoded image to the spectral images. Please start from the six channel reconstruction.
@inproceedings{chen2019computational,
title={A computational camera with programmable optics for snapshot high-resolution multispectral imaging},
author={Chen, Jieen and Hirsch, Michael and Eberhardt, Bernhard and Lensch, Hendrik PA},
booktitle={Computer Vision--ACCV 2018: 14th Asian Conference on Computer Vision, Perth, Australia, December 2--6, 2018, Revised Selected Papers, Part III 14},
pages={685--699},
year={2019},
organization={Springer}
}