A Review on Pedestrian Detection Techniques

Roshan Baba Kawade

Abstract: Pedestrian detection is a process in which number of person are detected from the image or the live video. This information send to user in the forms of sound or video alarm. It helps user in taking necessary preventive decision. Pedestrian detection process comes under security and safety aspect. As, security and safety is prime concern in todays digital world, it has gaining more concentration. It is being a very popular topic for research. This paper introduces various pedestrian detection techniques based on image processing and its various problems in detail. With variation in illumination, postures and clothing of pedestrian, detection process is a really difficult and challenging task. However, as demand of accuracy and speed of detection in real time is increasing, researchers are more focusing to tackle these problems. Efficient pedestrian detection with highest accuracy and speed of detection is required.

Keywords: Histogram of oriented gradient HOG, Support vector machine SVM, Advanced driver assistance system ADAS