Deep Learning Applied to Capacity Control in Commercial Establishments in Times of COVID-19

Rodriguez C.R.; Luque D.; Rosa C.L.; Esenarro D.; Pandey B. · Proceedings - 2020 12th International Conference on Computational Intelligence and Communication Networks, CICN 2020 (2020) · DOI: 10.1109/cicn49253.2020.9242584

Resumen

This research paper was developed to implement an intelligent solution for This research paper was developed to implement an intelligent solution for the control of the capacity of commercial establishments in times of COVID-19 using Yolo, which is a Convolutional Neural Network and a Deep Learning algorithm. For the application of this solution, a COCO dataset was used that is used in the implementation of Yolov4. A computer module was developed for the analysis of the flow of people, using Python 3.7, which mainly consists of an algorithm that determines the path and direction (movement) of a person, and this is evaluated in a limit o threshold that acts as the entrance and exit door of the main establishment; that is, it determines whether a person leaves or enters according to their route and direction. The results indicate that it is possible to implement this solution as an additional monitoring module for use as capacity control and with this offer a complete alternative to the owners of commercial establishments. In this way, it seeks to control the maximum capacity allowed due to the pandemic generated by the Sars-Cov.2 virus. The tests were conducted using an AMD Ryzen 7 3750H processor and an NVIDIA GTX 1660 TI video card. The possibility of determining whether the number of people who entered less than the number of people who left exceeds the maximum allowed by the pandemic on 50% of the real capacity.

Instituciones: Universidad Nacional Mayor de San Marcos; Universidad Nacional Federico Villarreal