Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/sofialtamsh/face-mask_detection

This project is a deep learning-based face mask detection system using Keras and TensorFlow. It employs a CNN classifier with ReLU activation to classify images into two categories: with mask and without mask, enabling real-time mask detection.
https://github.com/sofialtamsh/face-mask_detection

Last synced: 3 months ago
JSON representation

This project is a deep learning-based face mask detection system using Keras and TensorFlow. It employs a CNN classifier with ReLU activation to classify images into two categories: with mask and without mask, enabling real-time mask detection.

Awesome Lists containing this project

README 

        
# Face Mask Detection using Keras and TensorFlow 🧑‍⚕️🤖



## Problem Statement 🚨



In today’s world, ensuring safety and health in public places is crucial. While face masks have become a critical part of protecting ourselves and others, it can be difficult to monitor whether people are wearing them, especially in crowded environments. 



This project solves this problem by leveraging **deep learning** techniques to automatically detect whether individuals are wearing a face mask or not, helping in real-time monitoring for public safety in a variety of settings.



## Why Do We Need This? 💡



Face mask detection systems have become increasingly important as we look for ways to automate and streamline safety protocols. Here’s why this technology is needed:



- **Real-time Monitoring**: Automatically check if people are following mask-wearing protocols in crowded places like malls, public transport, offices, or airports.

- **Efficiency**: No need for manual supervision, which is time-consuming and often ineffective.

- **Scalability**: Easily scale up the detection system to handle hundreds or thousands of people with minimal infrastructure.

- **Cost-effective**: An automated system reduces the cost of human labor and provides instant alerts when someone isn’t wearing a mask.



## Advantages of This Project 🌟



- **Real-Time Detection**: The system uses a trained Convolutional Neural Network (CNN) to detect face masks in real-time.

- **High Accuracy**: With the help of Keras and TensorFlow, the system can achieve high accuracy in classifying images as "With Mask" or "Without Mask".

- **Scalable Solution**: Can be integrated into existing surveillance systems to provide an automated solution for mask detection in crowded spaces.

- **User-Friendly**: The solution is simple to implement and doesn't require deep technical knowledge to deploy.



## How It Works 🤔



### The Core Mechanism 🔍



At the heart of this project is a **Convolutional Neural Network (CNN)**, which is a type of deep learning model particularly well-suited for image classification tasks. This model analyzes the image pixels and detects whether an individual is wearing a face mask or not.



Here’s the detailed flow:



### 1. **Data Collection and Preprocessing 📊**



The first step involves gathering a labeled dataset consisting of images of people with and without masks. The dataset must be preprocessed to make it suitable for training:



- **Resizing** images to a standard size (e.g., 224x224).

- **Normalizing** pixel values to bring them within a range (usually between 0 and 1).

- **Data Augmentation** techniques, such as rotation, flipping, or zooming, are applied to increase the diversity of training data and avoid overfitting.



### 2. **Model Building with Keras and TensorFlow 🛠️**



Once the dataset is prepared, we build the **CNN model** using **Keras** (a high-level neural network API) and **TensorFlow** (the backend framework). 



The architecture of the CNN model consists of several key layers:



- **Convolutional Layers**: These layers apply filters to the image, detecting patterns such as edges, corners, or shapes. This is the most important step in learning image features.

- **Max Pooling Layers**: These layers help to reduce the spatial dimensions of the image, which reduces computation and helps prevent overfitting.

- **Fully Connected (Dense) Layers**: These layers take the features detected by the convolutional layers and make final predictions about whether the person in the image is wearing a mask.

- **Softmax Activation**: A softmax activation function is used in the output layer to classify the image into two categories: “With Mask” or “Without Mask.”



### 3. **Training the Model 🏋️‍♂️**



The model is trained using the labeled dataset. We use **categorical crossentropy** as the loss function, which is suitable for multi-class classification problems, and the **Adam optimizer** to minimize the loss function and adjust weights during training.



During training, the model adjusts its weights based on the error, gradually improving its ability to classify face mask images accurately.



### 4. **Model Evaluation 📈**



After training, the model is evaluated on a test dataset to check its performance. Key metrics include:



- **Accuracy**: Percentage of correct predictions (With Mask or Without Mask).

- **Precision & Recall**: These metrics are crucial when working with imbalanced datasets, ensuring the model correctly identifies masks without misclassifying.



### 5. **Real-Time Detection 🚶‍♂️**



Once trained, the model can be deployed for real-time face mask detection. The system uses a webcam or camera feed to continuously capture frames and then passes each frame through the trained model to classify whether the person is wearing a mask.



The result is displayed on the screen, and alerts can be sent if someone is detected without a mask.



## Why Use Python, TensorFlow, and Keras? 🐍



### **Python**: The Go-To Language for Machine Learning

Python is a powerful and widely used programming language for machine learning and deep learning projects. It offers a wide range of libraries (like TensorFlow and Keras) and has a large community, making it easy to find support and resources.



### **TensorFlow**: The Backbone for Deep Learning

TensorFlow is an open-source machine learning framework developed by Google. It’s highly scalable, supports both training and deployment of deep learning models, and has robust support for building and training neural networks.



### **Keras**: Simplifying Deep Learning

Keras, an easy-to-use interface for building deep learning models, is integrated with TensorFlow. Keras allows us to quickly prototype and experiment with neural network architectures without worrying about the low-level details.



## How to Run the Project 🏃‍♂️



To run this project, follow these simple steps:



1. **Clone the Repository**:

   ```bash

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

   cd face-mask-detection



## How to Run the Project 🏃‍♂️



To run this project, follow these simple steps:



2. **Install Dependencies**:

Ensure you have the required libraries installed by running:

```bash

pip install -r requirements.txt

```



3. **Prepare the Dataset**:

Place the images in the correct folders:

- `with_mask`: This folder should contain images of individuals wearing face masks.

- `without_mask`: This folder should contain images of individuals without face masks.



4. **Train the Model**:

Run the training script:

```bash

python train_model.py

```



5. ### Run Real-Time Detection:

After training, you can run the webcam detection:

```bash

python detect_mask.py

```



## Conclusion 🚀



This **Face Mask Detection** project is a powerful, deep learning-based solution that leverages the capabilities of **Keras** and **TensorFlow** to automatically detect face masks in real-time. With the growing need for safety and monitoring, this project can be applied in a wide range of environments, offering both efficiency and scalability.



By utilizing **convolutional neural networks**, **Python**, and modern deep learning frameworks, this system can be deployed in various industries, helping to ensure that health and safety protocols are followed effectively.