face-mask-detection-system
😷 Face Mask Detection System
A deep learning-based Face Mask Detection System that identifies whether a person is wearing a mask or not using image input and real-time webcam.
Built using MobileNetV2 (Transfer Learning) and OpenCV DNN face detection.
🚀 Features
- 📸 Mask detection from uploaded images
- 🎥 Real-time detection using webcam
- 🧠 Transfer learning with MobileNetV2
- 📊 Accuracy and loss visualization
- ⚡ Fast and efficient performance
🎥 Demo
🟢 With Mask Detection
🔴 Without Mask Detection
🌐 Streamlit Web App Interface
🛠️ Tech Stack
- Python
- TensorFlow / Keras
- OpenCV
- Streamlit
- NumPy, Matplotlib, Scikit-learn
📂 Project Structure
```bash id=”p4u2hk” face-mask-detection-system/ │── app.py │── detect_realtime.py │── train_model.py │── requirements.txt │── README.md │── mask_detector.h5 │── deploy.prototxt │── res10_300x300_ssd_iter_140000.caffemodel │── screenshots/ │── .gitignore │── LICENSE
---
## 📊 Model Performance
* Training Accuracy: ~94%
* Validation Accuracy: ~96%
---
## 📁 Dataset
Dataset used: **Face Mask Detection Dataset (Kaggle)**
* Total Images: 7553
* Classes:
* 😷 With Mask
* ❌ Without Mask
👉 https://www.kaggle.com/datasets/omkargurav/face-mask-dataset
---
## ⚙️ Installation
```bash id="pj0wya"
git clone https://github.com/shanmukha0527/face-mask-detection-system.git
cd face-mask-detection-system
pip install -r requirements.txt
▶️ Usage
🔹 Run Streamlit Web App
```bash id=”czkmbx” streamlit run app.py
Upload image → get prediction instantly.
---
### 🔹 Run Real-Time Detection
```bash id="24i45z"
python detect_realtime.py
Press Q to exit.
🔹 Train Model (Optional)
bash id="4e4xdt"
python train_model.py
⚡ How It Works
- Detect face using OpenCV DNN
- Extract face region
- Resize to 224×224
- MobileNetV2 predicts mask / no mask
- Display result with bounding box
⚠️ Important Notes
- ❌ Do NOT upload
kaggle.json - Dataset is not included in repo
- Model file (
.h5) is pre-trained
📌 Future Improvements
- 🔔 Alert system for no-mask detection
- 📊 Count number of people without mask
- 🌐 Deploy on Streamlit Cloud
- 📱 UI improvements
👨💻 Author
Gondrala Shanmukha Akhilesh B.Tech CSE (AI & ML)