Face Mask Detection Model

demo: https://face-mask-detection-using-cnn-1.onrender.com

A deep learning model for detecting whether a person is wearing a face mask or not, built with PyTorch.

The system can:

Classify single images

Process batches of images from a directory

Perform real-time face mask detection using webcam

Detect multiple faces in an image and classify each one

The model achieves high accuracy in distinguishing between "with_mask" and "without_mask" classes.

Model Architecture

The model is a custom convolutional neural network (CNN) with the following architecture:

Feature Extractor

• ConvBlock 1: 3→32 channels, MaxPool (128×128→64×64)
• ConvBlock 2: 32→64 channels, MaxPool (64×64→32×32)
• ConvBlock 3: 64→128 channels, MaxPool (32×32→16×16)
• ConvBlock 4: 128→256 channels, MaxPool (16×16→8×8)
• ConvBlock 5: 256→512 channels, MaxPool (8×8→4×4)

Classifier

• Fully Connected 1: 8192→512 units with Dropout (0.5)
• Fully Connected 2: 512→2 units (with_mask, without_mask)

face_mask_video.mp4.mp4

Performance

• Test Accuracy: 98.35%
• Training Time: ~25 epochs
• Input Size: 128×128 RGB images

Installation

1. Clone the repository:

git clone https://github.com/olyadboka/face-mask-detection.git
cd face-mask-detection

2. Install dependencies:

pip install -r requirements.txt

3. to Execute

   python test.py --camera
   or
   python test.py --image image1.jpeg