Face mask detection

Face mask detection on images and videos.

People wearing/not-wearing facemasks are detected and tracked over the video.

Description

Two-model approach:

• Face detection model.
• Face mask detection with the first model as an input.

CNN architecture:

• Image input.
• Convolutional layer: VGG net.
• Batch Normalization.
• Pooling
• Output

LSTM for tracking and location prediction:

• Input sequence
• Visual features
• Spatial information
• Temporal information
• Prediction