Title :
Meeting in the Middle: A top-down and bottom-up approach to detect pedestrians
Author :
Shaukat, Arslan ; Gilbert, A. ; Windridge, David ; Bowden, Richard
Author_Institution :
CVSSP, Univ. of Surrey, Guildford, UK
Abstract :
This paper proposes a generic approach combining a bottom-up (low-level) visual detector with a top-down (high-level) fuzzy first-order logic (FOL) reasoning framework in order to detect pedestrians from a moving vehicle. Detections from the low-level visual corner based detector are fed into the logical reasoning framework as logical facts. A set of FOL clauses utilising fuzzy predicates with piecewise linear continuous membership functions associates a fuzzy confidence (a degree-of-truth) to each detector input. Detections associated with lower confidence functions are deemed as false positives and blanked out, thus adding top-down constraints based on global logical consistency of detections. We employ a state of the art visual detector on a challenging pedestrian detection dataset, and demonstrate an increase in detection performance when used in a framework that combines bottom-up detections with (fuzzy FOL-based) top-down constraints.
Keywords :
fuzzy logic; fuzzy set theory; image motion analysis; inference mechanisms; object detection; pedestrians; piecewise linear techniques; road vehicles; traffic engineering computing; FOL clause set; bottom-up approach; bottom-up detections; bottom-up visual detector; false positives; fuzzy FOL-based top-down constraints; fuzzy confidence; fuzzy predicates; global logical consistency; logical facts; low-level visual corner based detector; lower confidence functions; moving vehicle; pedestrian detection dataset; piecewise linear continuous membership functions; state of the art visual detector; top-down approach; top-down fuzzy first-order logic reasoning framework; Cognition; Data mining; Detectors; Feature extraction; Fuzzy logic; Vehicles; Visualization;
Conference_Titel :
Pattern Recognition (ICPR), 2012 21st International Conference on
Conference_Location :
Tsukuba
Print_ISBN :
978-1-4673-2216-4