Learning joint intensity-depth sparse representations

Ivana Tosic, Sarah Drewes

This paper presents a method for learning overcomplete dictionaries composed of two modalities that describe a 3D scene: image intensity and scene depth. We propose a novel Joint B…