A Framework to Enhance Object Detection Performance by using YOLO Algorithm
R. Shiva Shankar, L V Srinivas, P Neelima, Gadiraju Mahesh
Abstract
One of the most widely used applications in computer vision is object detection. It is a technique used for detecting and sculpting real-world objects. Despite the fact that there are numerous detection methods, their accuracy and efficiency are insufficient. This research work utilizes deep learning approaches to recognise objects using the YOLOv3 and YOLOv4 algorithms. The images of the two classes, Auto and Street Light, were combined to produce a dataset. Both training and testing are done using this dataset. This algorithm can recognise several things quickly and accurately, making it one of the most efficient methods for detecting a variety of objects. Various existing models are contrasted with Assessment methods such as Precision, Recall, F1-Score, average IOU, mean Average Precision (mAP), and others in the context of Object Detection.