Object detection using CNN and YoloV5

Abstract A YOLOv5 model-based object detection software is presented in this paper. A number of forms of media sources, such as pictures, videos, and real-time videos, can be used with the program. Users can choose the location of the model weights, the media to be analyzed, the size of the image processing to be done, and more. After identifying items in the media, the script outputs those objects together with the bounding boxes that belong to them. It uses a method known as Non-Max Suppression (NMS) to get rid of duplicate detections and provides options for saving the results in several formats or seeing them right on the screen. For possibly better detection accuracy over time, the script can optionally be set up to update the YOLOv5 model weights.