YOLO v5 and Faster R-CNN Performance Evaluation of Solid Waste in Object Detection Application

Abstract

Object detection is a booming technology that is on par with computer vision and image processing in which an object of a specific type is detected in an image or video. Object detection consists of several approaches like Retina-Net, Single Shot MultiBox Detector (SSD), and Faster R-CNN. These approaches are used in object detection with limited data, but these approaches either run in two algorithms or has high execution time; to overcome these limitations, the authors have used the latest version of Yolo with the custom dataset of solid waste. In this algorithm, an image in the solid waste dataset, which was annotated, labelled, pre-processed, and segmented and a build version is created with the yolo model; this version can either be used directly in the code for online execution or downloaded in the local system for offline execution.