Fine-tuned a MobileNetV2 CNN architecture to identify different facial landmarks and used it to detect ROI, along with a pre-trained neural network (SSD). Integrated alert sound mode to produce sound if someone is not wearing a face mask, all in real-time.