Comparative Analysis of Novel Social Distancing Detector using Enhanced You Only Look Once and CNN Deep Learning Approach with improved Accuracy

Abstract

The objective is detecting Novel Social Distancing using CNN techniques in comparison with YOLO. Materials and Methods: Social Distance Detector is implemented using two deep learning algorithms, CNN techniques (N=20) and YOLO (N=20) algorithms. YOLO is an effective and efficient real time algorithm used for object detection. Convolutional Neural Networks is a deep learning algorithm where neural networks are used for image recognition and classification. Google AI open Images dataset is used for image detection. Dataset contains more than 10,000 images. Every set 20 samples are considered and splitted to testing and training sets. Results and Discussion: Accuracy of YOLO algorithm is 96.26% and CNN techniques is 91.28%. The statistical varied value between YOLO algorithm next CNN techniques with (p<0.05). Conclusion: YOLO algorithm proves best than CNN techniques for Novel Social Distancing Detection.