Visual Localization through CBIR with RGB+D data

(3D Bild-Analyse und Synthese)
Betreuer: Anas Al-Nuaimi
In this thesis, the student will extend the methods developed at LMT to perform place recognition using a range image-based enviroment representation and concepts from Content-based Image Retrieval (CBIR). Specifically different fusion approaches are to be considered that allow joint RGB+D retrieval for improved location recognition performance. This thesis is an extension of a previous with the same title.
Anforderungen: Computer Vision. Image Processing