Towards using smartphones as hyperspectral cameras

Spectral reconstruction (SR, recovering spectra from RGB measurements) is a vital problem of computational photography. However, even recent algorithms do not provide acceptable accuracy. As a matter of curiosity, modern mobile devices open a new opportunity to improve the quality of SR by utilizing images from several cameras at once. This leads to the idea of creating a mobile hyperspectral camera for the general public. In this paper we investigate the achievable accuracy when using several identical cameras simultaneously in combination with different spectral filters. To find optimal filters, two algorithms are proposed: one learns linear regression for SR and the other learns fully connected neural network. As a result of numerical experiments, it is found that in the case of 4 cameras and 4 filters, the SR accuracy is three times higher than in the case of SR from RGB images.