Detection And Tracking of Multiple Pedestrians Using Deep Learning
Ayush Kumar, Tripty Singh, Prakash Duraisamy · 2023
Recent advancements in deep learning and computer vision have benefitted security systems. Recent object tracking algorithms, in example, have incorporated deep learning in a variety of methods to enhance tracking performance. Object tracking and security systems continue to provide a number of difficulties. This work will first identify the several challenges related to the object tracking for an autonomous system and further the object tracker system will be evaluated in a simulated environment for multiple objects like pedestrian, moving or stationary object. This research paper is all about designing the tracking algorithm for multiple pedestrians. This research employed YOLO V3 with a Deep Sort architecture. In order to get heightened accuracy different kind of scenario is used for preparing the test data. As a result, precision is 93 %, Recall is 98% and accuracy is 96%.