Semantic Mapping Using Object-Class Segmentation of RGB-D Images

By Nenad Biresev | Autonomous Intelligent Systems | March 2012
Download thesis

The thesis describes a novel approach to recognizing objects in RGB-D images and for making this information persistent in a 3D semantic map. It makes two major contributions: Firstly, he proposes a novel approach to object-class segmentation in RGB-D images based on random forest classifiers that provides a pixel-wise class labeling in the image. 

Secondly, he proposes a principled method to fuse image segmentations from multiple viewpoints in a 3D semantic map. To this end, he integrated an approach to simultaneous localization and mapping based on Multiresolution Surfel maps. Experiments show that Bayesian fusion in 3D improves on the recognition rate significantly, compared to recognition from individual views.
Nenad Biresev was received the 2012 NRW Young Scientist Award for the IROS paper summarizing the results of the thesis.

Please publish modules in offcanvas position.

Free Joomla! template by L.THEME