The application of convolutional neural networks for tomographic reconstruction of hyperspectral images

A novel method, utilizing convolutional neural networks (CNNs), is proposed to reconstruct hyperspectral cubes from computed tomography imaging spectrometer (CTIS) images. Current reconstruction algorithms are usually subject long times and mediocre precision in cases of a large number spectral channels. The constructed CNNs deliver higher shorter time than sparse expectation maximization algorithm. In addition, the network can handle two different types real-world images at same time—specifically ColorChecker carrot considered. This work paves way toward real-time CTIS