Comparative study of pedestrian detection techniques for driver assistance system

Pedestrian detection is one of the major concerns in the field of intelligent transportation systems. Recently, the deep learning based algorithms has greatly improved the accuracy of pedestrian detection. The major challenge in pedestrian detection is varying body shape, posture and different color of clothes. This article compares the performance of YOLO v3 and HOG SVM model on INRIA person dataset containing pedestrians in different pose, shape and lighting conditions. Bounding boxes are generated to indicate the location of the pedestrian in an input image. The experimentation results obtained are presented.