Mask-Detector

This repo contains a training setup for creating a model that can detect faces and label them as wearing\not wearing masks.

---

Settings

• Framework: PyTorch
• Dataset: Custom build, by generating images with mask wearing faces from empty faces. (Used a modified script from https://github.com/prajnasb/observations)
• Faster RCNN Architecture with pretrained MobileNet v2 backbone
• To train: python train_run_builder.py (to change the training setup edit config.py)
• To run the model on an image or on video stream from a local camera: python detect.py --realtime=False\True

COCO Metrics

IoU metric: bbox

Score	IOU	AREA	RESULTS
Average Precision	(AP) @[ IoU=0.50:0.95	area= all	maxDets=100 ] = 0.640
Average Precision	(AP) @[ IoU=0.50	area= all	maxDets=100 ] = 0.990
Average Precision	(AP) @[ IoU=0.75	area= all	maxDets=100 ] = 0.799
Average Precision	(AP) @[ IoU=0.50:0.95	area= small	maxDets=100 ] = -1.000
Average Precision	(AP) @[ IoU=0.50:0.95	area=medium	maxDets=100 ] = 0.609
Average Precision	(AP) @[ IoU=0.50:0.95	area= large	maxDets=100 ] = 0.653
Average Recall	(AR) @[ IoU=0.50:0.95	area= all	maxDets= 1 ] = 0.693
Average Recall	(AR) @[ IoU=0.50:0.95	area= all	maxDets= 10 ] = 0.693
Average Recall	(AR) @[ IoU=0.50:0.95	area= all	maxDets=100 ] = 0.693
Average Recall	(AR) @[ IoU=0.50:0.95	area= small	maxDets=100 ] = -1.000
Average Recall	(AR) @[ IoU=0.50:0.95	area=medium	maxDets=100 ] = 0.650
Average Recall	(AR) @[ IoU=0.50:0.95	area= large	maxDets=100 ] = 0.706