rkshiyaniya / Real-Time-Face-Mask-Detector

Real Time Face Mask Detector with TensorFlow (Keras), OpenCV. By Caffe model detecting Face then Classify whether person wearing Mask or Not based on trained model on base of mobilenet_v2 model with pre-trained weights of 'imagenet'.

Real Time Face Mask Detector using Deep Learning

Real Time Face Mask Detector with TensorFlow/Keras with the help of OpenCV. By Caffe model detecting Face then Classify whether person wearing Mask or Not based on trained model on base of mobilenet_v2 model with pre-trained weights of 'imagenet'.

Requirements

• TensorFlow
• Python 3.5 +
• Numpy
• Opencv
• imutils
• Matplotlib
• sklearn

Note

• The dataset used can be downloaded here - Click to Download
• Caffe-based Face Detector because it's more accurate and fast
• Trained model on base of mobilenet_v2 model with pre-trained weights of 'imagenet'

About Project

• First Load Data and Pre-process it

• Trained model on base of mobilenet_v2 model with pre-trained weights of 'imagenet'

• Save model as 'model_detector.model' ( save_format = h5 )

• Using OpenCV and serialized faceNet (res10_caffe_model) extract Person's Live Face and it's location

• Then our trained serialized 'mask_detector.model' detect whether person wearing mask or not

• Then rest of the work will be done by OpenCV as to label prediction and show live

Screenshots

• With Mask ...

• Without Mask ...

Accuracy Plot