Project Overview

We will create a system that detects in real-time whether a person visible on a webcam is wearing a mask or not. The model will be trained using Keras and OpenCV.

Dataset

The dataset contains 1376 images, with 690 images of people wearing masks and 686 of people without masks.

Tools Required

We will use Jupyter Notebook for development. After installing it using:

bash

pip3 install notebook

You can run the server by typing:

bash

jupyter notebook

This will open Jupyter in your browser. Create a new project and write your code.

Project Structure

We’ll build the project in two parts:

Train a Face Mask Detector model using Keras.

Test the model in real-time using a webcam with OpenCV.

Part 1: Model Training (train.py)

1. Import Required Libraries

python

from keras.optimizers import RMSprop from keras.preprocessing.image import ImageDataGenerator import cv2 from keras.models import Sequential from keras.layers import Conv2D, MaxPooling2D, Flatten, Dropout, Dense from keras.callbacks import ModelCheckpoint import numpy as np

2. Build CNN Model

python

model = Sequential([    Conv2D(100, (3,3), activation='relu', input_shape=(150, 150, 3)),    MaxPooling2D(2,2),    Conv2D(100, (3,3), activation='relu'),    MaxPooling2D(2,2),    Flatten(),    Dropout(0.5),    Dense(50, activation='relu'),    Dense(2, activation='softmax') ]) model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['acc'])

3. Data Augmentation

python

train_datagen = ImageDataGenerator(rescale=1.0/255,                                   rotation_range=40,                                   width_shift_range=0.2,                                   height_shift_range=0.2,                                   shear_range=0.2,                                   zoom_range=0.2,                                   horizontal_flip=True,                                   fill_mode='nearest') train_generator = train_datagen.flow_from_directory('./train', batch_size=10, target_size=(150, 150)) validation_datagen = ImageDataGenerator(rescale=1.0/255) validation_generator = validation_datagen.flow_from_directory('./test', batch_size=10, target_size=(150, 150))

4. Save Best Model

python

checkpoint = ModelCheckpoint('model2-{epoch:03d}.model', monitor='val_loss', save_best_only=True, mode='auto')

5. Train the Model

python

history = model.fit_generator(train_generator,                              epochs=10,                              validation_data=validation_generator,                              callbacks=[checkpoint])

Part 2: Real-Time Mask Detection (test.py)

python

import cv2 import numpy as np from keras.models import load_model model = load_model("./model-010.h5") results = {0: 'without mask', 1: 'mask'} GR_dict = {0: (0, 0, 255), 1: (0, 255, 0)} rect_size = 4 cap = cv2.VideoCapture(0) haarcascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml') while True:    (rval, im) = cap.read()    im = cv2.flip(im, 1, 1)    rerect_size = cv2.resize(im, (im.shape[1] // rect_size, im.shape[0] // rect_size))    faces = haarcascade.detectMultiScale(rerect_size)    for f in faces:        (x, y, w, h) = [v * rect_size for v in f]        face_img = im[y:y+h, x:x+w]        rerect_sized = cv2.resize(face_img, (150, 150))        normalized = rerect_sized / 255.0        reshaped = np.reshape(normalized, (1, 150, 150, 3))        result = model.predict(reshaped)        label = np.argmax(result, axis=1)[0]        cv2.rectangle(im, (x, y), (x+w, y+h), GR_dict[label], 2)        cv2.rectangle(im, (x, y-40), (x+w, y), GR_dict[label], -1)        cv2.putText(im, results[label], (x, y-10), cv2.FONT_HERSHEY_SIMPLEX, 0.8, (255, 255, 255), 2)    cv2.imshow('LIVE', im)    key = cv2.waitKey(10)    if key == 27:        break cap.release() cv2.destroyAllWindows()

Run this script using:

bash

python3 test.py

Summary

In this project, we built a real-time Face Mask Detector using Python, Keras, and OpenCV. The project involved training a convolutional neural network to detect whether a person is wearing a mask, followed by deploying the model to work with a live webcam feed.