Title :
Pedestrian detection - a neuromorphic approach
Author :
Woo Joon Han ; Il Song Han
Author_Institution :
Korea Adv. Inst. of Sci. & Technol., Daejeon, South Korea
Abstract :
In this paper, an algorithm inspired from human brain was implemented. The characteristics of visual cortex were used to extract orientation features from an image taken from a hand held camera on a bus. The result showed detection of pedestrians nearer to the camera so to ignore pedestrians that are too far away for any danger in pedestrian detection for vehicle application. Also the detection of pedestrian with bigger motion that represents danger for a vehicle was tested with success the idea that human eye has more interest in objects in motion rather than the image in whole. The feasibility of applying this idea is further extended when it can be implemented electronically using 0.18um CMOS technology.
Keywords :
CMOS integrated circuits; cameras; feature extraction; image motion analysis; object detection; road vehicles; traffic engineering computing; CMOS technology; feature extraction; hand held camera; human brain; human eye; neuromorphic approach; pedestrian detection; pedestrian motion; size 0.18 mum; vehicle application; visual cortex; Image recognition; Visualization;
Conference_Titel :
Circuits and Systems (MWSCAS), 2011 IEEE 54th International Midwest Symposium on
Conference_Location :
Seoul
Print_ISBN :
978-1-61284-856-3
Electronic_ISBN :
1548-3746
DOI :
10.1109/MWSCAS.2011.6026529
