From Pattern Recognition to Place Identication
Sven Eberhardt, Tobias Kluth, Christoph Zetzsche, Kerstin Schill · 2012
homepage: http://www.informatik.uni-bremen.de/cog neuroinf/en Abstract. What are the ingredients required for vision-based place recog- nition? Pattern recognition models for localization must fulll invariance requirements dierent from those of object recognition. We propose a method to evaluate the suitability of existing image processing tech- niques by testing their outputs against these invariances. The method is applied to several holistic and one local model. We generalize our nd- ings and identify model properties of locality, spatial conguration and