Wearing masks is legally mandated in numerous countries globally, and there is a growing trend where private entities in those regions are also adopting this practice. It can be challenging to monitor compliance with these essential social distance standards in large settings. While vaccines are available, social distance and wearing masks are considered appropriate rules besides the spread of the pandemic corona family infection in the current situation. Our paper's objective is to develop a deep learning framework that facilitates both social distance monitoring and face mask recognition systems. A self-making calibration video tests it from a camera. Now recognize the humans in video sequences using the YOLOv5 pre-train human tracking model, and in this model, social distance tracking with rectangle bounding boxes, and again, a new YOLOv5 bird's eye view approach used with circular bounding boxes. Our face mask action recognition model has achieved 97.5% mean average precision. Now results show that the recognized system effectively identifies individuals who violate public distances.