Social Distance Detection Method Using Machine Learning Algorithm in Covid-19

Authors

  • Priyanka Tyagi, NeelamShrivastava, Sudeshna Chakrabory

Keywords:

Machine Learning, Deep Learning, You Only Look Once, Region-based Convolution Neural Networks, Multi-object tracking, Multi-Type Tree

Abstract

The motivation in the back of this paper is, consequently, to present a profound gaining knowledge of stage to social distance following using an upward point of view. The shape utilizes the YOLOv3 object acknowledgment worldview to distinguish people in video successions. The exchange studying method is additionally executed to increase the exactness of the version. To assess social distance infringement between individuals, we applied a guess of actual distance to pixel and set an part. An infringement restriction is laid out to evaluate whether or now not the space esteem penetrates the bottom social1distance aspect. Moreover, the following calculation is utilized to differentiate people in video preparations to such an extent that the person who disregards/passes the social1distance boundary is likewise being observed. Tests are finished on diverse video groupings to test the productivity of the version. The precision of 92% and 98% completed with the aid of the popularity model with out and with pass studying, personally. The following precision of the model is 95%.

Downloads

Published

2022-08-16