Multi-scale Bag-Of-Features for large-size map retrieval

Retrieving a large collection of environment maps built by mapper robots is a key problem for mobile robot self-localization. The current paper studies this map retrieval problem from a novel perspective of a multi-scale Bag-Of-Features (BOF) approach. In general, multi-scale approach is advantageous in capturing both the global structure and the local details of a given map. On the other hand, BOF map retrieval is advantageous in its compact map representation as well as efficient map retrieval using an inverted file system. Combining the advantages of both approaches is the main contribution of this paper. Our approach is based on multi-cue BOF as well as BOF dimension reduction, and achieves efficiency and compactness of the map retrieval system. Experiments on a large collection of point feature maps show promising results.