Tauqueer Alam

View Demo

Mask Detection Using Deep Learning: A Step Towards Safer Environments

Introduction

In the wake of global pandemics, ensuring public safety through mask-wearing has become crucial. With the power of Deep Learning and Computer Vision, we can automate mask detection using Neural Networks and Convolutional Neural Networks (CNNs).

This project detects whether a person is wearing a mask or not in real-time using a Deep Learning model trained with CNN, achieving 99% accuracy. The model classifies images into three categories:

Technologies Used

Model Training Process

1. Dataset Preparation

Used a dataset containing images of people with and without masks. Applied data augmentation to increase dataset diversity.

2. Model Architecture

The model follows a CNN-based architecture with layers including:

Activation functions: ReLU for hidden layers, Softmax for classification.

3. Training and Accuracy

Model trained with Adam optimizer and Categorical Cross-Entropy loss, achieving an impressive 99% accuracy on the test set.

Deployment with Flask

Created a Flask web app to deploy the model. Integrated OpenCV for real-time video stream processing. The app captures live video from the webcam and classifies faces on the screen.

Results & Future Improvements

Conclusion

This Mask Detection project demonstrates how Deep Learning can be used for real-time safety enforcement. By leveraging CNNs and Flask, we can deploy an intelligent system that helps ensure public health and safety.

Would love to hear your thoughts! Feel free to check out the code and contribute to further improvements. 🚀

You can check out the full project on my GitHub or view a live demo below:

View Demo