Colliding Bodies Optimization With Deep Belief Network Based Robust Pedestrian Detection

Pedestrian detection is a significant research topic in the computer vision (CV) domain for longer period. Recently, deep learning (DL) and specifically convolutional neural network (CNN) exhibit improvement tasks such as object detection, segmentation, image classification, etc. With this motivation, study develops novel Colliding Bodies Optimization with Deep Learning based Robust Detection (CBODL-RPD) model. The goal of CBODL-RPD approach to identify occurrence pedestrians non-pedestrians via process. For process, YOLO v4 Adagrad optimizer applied. In addition, technique employs SqueezeNet model generate feature vectors, hyperparameter tuning process performed CBO algorithm. At last, belief (DBN) applied accurate pedestrian detection. A comprehensive experimental analysis made demonstrate results technique. comparative outcome reported improved outcomes method over other recent methods.