This study focuses on examining the performance of state-of-the-art detection algorithms to detect various makeshift masks used by Indians. Our research can be used to implement efficient AI entrance monitoring systems to automate the entry of staff, customers and students in workplaces and institutions.
We've experimented with 3 real-time object detection pipelines, namely - SSD MobileNetV2, YOLO v3 and YOLO v4.