COVID-19: Face Mask Detection using TensorFlow and OpenCV

栏目: IT技术 · 发布时间: 4年前

内容简介:Construct a CNN model to detect if a person is wearing a face mask or not with your webcam or mobile camera.In these toughFor building this model, I will be using the face mask dataset provided by

Construct a CNN model to detect if a person is wearing a face mask or not with your webcam or mobile camera.

In these tough COVID-19 times, wouldn’t it be satisfying to do something related to it? I decided to build a very simple and basic Convolutional Neural Network (CNN) model using TensorFlow with Keras library and OpenCV to detect if you are wearing a face mask to protect yourself. Interesting! Isn’t it?

COVID-19: Face Mask Detection using TensorFlow and OpenCV

Photo by Macau Photo Agency on Unsplash

For building this model, I will be using the face mask dataset provided by Prajna Bhandary . It consists of about 1,376 images with 690 images containing people with face masks and 686 images containing people without face masks.

I am going to use these images to build a CNN model using TensorFlow to detect if you are wearing a face mask by using the webcam of your PC . Additionally, you can also use your phone’s camera to do the same!

Step 1: Data Visualization

In the first step, let us visualize the total number of images in our dataset in both categories. We can see that there are 690 images in the ‘ yes ’ class and 686 images in the ‘ no ’ class.

The number of images with facemask labelled 'yes': 690 
The number of images with facemask labelled 'no': 686

Step 2: Data Augmentation

In the next step, we augment our dataset to include more number of images for our training. In this step of data augmentation , we rotate and flip each of the images in our dataset. We see that, after data augmentation, we have a total of 2751 images with 1380 images in the ‘ yes ’ class and ‘ 1371 ’ images in the ‘ no ’ class.

Number of examples: 2751 
Percentage of positive examples: 50.163576881134134%, number of pos examples: 1380 
Percentage of negative examples: 49.836423118865866%, number of neg examples: 1371

Step 3: Splitting the data

In this step, we split our data into the training set which will contain the images on which the CNN model will be trained and the test set with the images on which our model will be tested.

In this, we take split_size =0.8 , which means that 80% of the total images will go to the training set and the remaining 20% of the images will go to the test set.

The number of images with facemask in the training set labelled 'yes': 1104
The number of images with facemask in the test set labelled 'yes': 276
The number of images without facemask in the training set labelled 'no': 1096
The number of images without facemask in the test set labelled 'no': 275

After splitting, we see that the desired percentage of images have been distributed to both the training set and the test set as mentioned above.

Step 4: Building the Model

In the next step, we build our Sequential CNN model with various layers such as Conv2D, MaxPooling2D, Flatten, Dropout and Dense . In the last Dense layer, we use the ‘ softmax ’ function to output a vector that gives the probability of each of the two classes.

Sequential Model for Face Mask detection

Here, we use the ‘ adam ’ optimizer and ‘ binary_crossentropy ’ as our loss function as there are only two classes. Additionally, you can even use the MobileNetV2 for better accuracy.

Step 5: Pre-Training the CNN model

After building our model, let us create the ‘ train_generator ’ and ‘ validation_generator ’ to fit them to our model in the next step. We see that there are a total of 2200 images in the training set and 551 images in the test set .

Found 2200 images belonging to 2 classes. 
Found 551 images belonging to 2 classes.

Step 6: Training the CNN model

This step is the main step where we fit our images in the training set and the test set to our Sequential model we built using keras library. I have trained the model for 30 epochs (iterations). However, we can train for more number of epochs to attain higher accuracy lest there occurs over-fitting .

history = model.fit_generator(train_generator,
                              epochs=30,
                              validation_data=validation_generator,
                              callbacks=[checkpoint])>>Epoch 30/30
220/220 [==============================] - 231s 1s/step - loss: 0.0368 - acc: 0.9886 - val_loss: 0.1072 - val_acc: 0.9619

We see that after the 30th epoch, our model has an accuracy of 98.86% with the training set and an accuracy of 96.19% with the test set. This implies that it is well trained without any over-fitting.

Step 7: Labeling the Information

After building the model, we label two probabilities for our results. [‘0’ as ‘without_mask’ and ‘1’ as ‘with_mask’] . I am also setting the boundary rectangle color using the RGB values. [‘RED’ for ‘without_mask’ and ‘GREEN’ for ‘with_mask]

labels_dict={0:'without_mask',1:'with_mask'} 
color_dict={0:(0,0,255),1:(0,255,0)}

Step 8: Importing the Face detection Program

After this, we intend to use it to detect if we are wearing a face mask using our PC’s webcam. For this, first, we need to implement face detection. In this, I am using the Haar Feature-based Cascade Classifiers for detecting the features of the face.

face_clsfr=cv2.CascadeClassifier('haarcascade_frontalface_default.xml')

This cascade classifier is designed by OpenCV to detect the frontal face by training thousands of images. The .xml file for the same needs to be downloaded and used in detecting the face. I have uploaded the file in my GitHub repository.

Step 9: Detecting the Faces with and without Masks

In the last step, we use the OpenCV library to run an infinite loop to use our web camera in which we detect the face using the Cascade Classifier . The code webcam = cv2.VideoCapture(0) denotes the usage of webcam.

The model will predict the possibility of each of the two classes ([without_mask, with_mask]) . Based on which probability is higher, the label will be chosen and displayed around our faces.

Additionally, you can download the DroidCam application for both Mobile and PC to use your mobile’s camera and change the value from 0 to 1 in webcam= cv2.VideoCapture(1).

Demo

Let us see the demo where I try it out on myself!

COVID-19: Face Mask Detection using TensorFlow and OpenCV

Demo Video

From the above demo video, we see that the model was correctly able to detect if I was wearing a mask or not and displays the same on the label.

I am sharing the link to my GitHub repository where you can find the detailed .ipynb program code for you to try it out.

In this article, we have successfully built a CNN model to detect if a person is wearing a face mask or not. This can be used in numerous applications. Sporting a mask may be necessary in the near future, considering the COVID-19 crisis and this method to detect if the person wears a face mask may come in handy.

I do hope that these tough times come to an end soon. Let us all stay healthy and be safe. Till then, happy machine learning!


以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持 码农网

查看所有标签

猜你喜欢:

本站部分资源来源于网络,本站转载出于传递更多信息之目的,版权归原作者或者来源机构所有,如转载稿涉及版权问题,请联系我们

谷歌的断舍离:互联网企业的破坏式创新

谷歌的断舍离:互联网企业的破坏式创新

[日]辻野晃一郎 / 樊颖 / 机械工业出版社 / 2018-1 / 45.00

本书主要分为三部分: 第一部分主要讨论了世界当下如火如荼的互联网企业进军传统产业大潮,并探讨了传统企业在互联网时代的救赎之路。 第二部分主要探讨了成功体验的反面:速度与迭代,并讨论了传统企业之所以无法实现迭代与快速发展的关键原因。介绍互联网公司如何通过组织精简流程来实现高速竞争时代的机动性。 第三部分讨论了互联网时代究竟需要什么样的人才,传统企业的员工应当怎样投身互联网企业才能避......一起来看看 《谷歌的断舍离:互联网企业的破坏式创新》 这本书的介绍吧!

JS 压缩/解压工具
JS 压缩/解压工具

在线压缩/解压 JS 代码

JSON 在线解析
JSON 在线解析

在线 JSON 格式化工具

URL 编码/解码
URL 编码/解码

URL 编码/解码