Deep Hyperspectral Prior: Denoising, Inpainting, Super-Resolution
This addresses the challenge of limited training data in hyperspectral image processing, though it is an incremental extension of existing deep prior methods to a new domain.
The paper tackles the problem of hyperspectral image restoration tasks like denoising, inpainting, and super-resolution, where datasets are small, by proposing a deep prior approach using CNNs without training, achieving performance comparable to trained networks.
Deep learning algorithms have demonstrated state-of-the-art performance in various tasks of image restoration. This was made possible through the ability of CNNs to learn from large exemplar sets. However, the latter becomes an issue for hyperspectral image processing where datasets commonly consist of just a few images. In this work, we propose a new approach to denoising, inpainting, and super-resolution of hyperspectral image data using intrinsic properties of a CNN without any training. The performance of the given algorithm is shown to be comparable to the performance of trained networks, while its application is not restricted by the availability of training data. This work is an extension of original "deep prior" algorithm to HSI domain and 3D-convolutional networks.