Programmable Spectrometry -- Per-pixel Classification of Materials using Learned Spectral Filters

Many materials have distinct spectral profiles. This facilitates estimation of the material composition of a scene at each pixel by first acquiring its hyperspectral image, and subsequently filtering it using a bank of spectral profiles. This process is inherently wasteful since only a set of linear...

Popoln opis

Bibliografske podrobnosti

Autores principales:	Saragadam, Vishwanath, Sankaranarayanan, Aswin C.
Format:	Artículo
Izdano:	2019
Teme:	Image and Video Processing Computer Vision and Pattern Recognition Machine Learning
Online dostop:	https://arxiv.org/abs/1905.04815 http://arxiv.org/abs/1905.04815
Oznake:	Označite Brez oznak, prvi označite!