MyAssignmentServices uses cookies to deliver the best experience possible. Read more

  • Internal Code :
  • Subject Code :
  • University :
  • Subject Name : IT Computer Science

Real-Time Face Mask Detection and Face Recognition Using CapsNet Technique

Research Description:

At the start of this year 2020 everything was normal till the corona virus hit for the very first time in world. Especially in china from where it emerges. Soon after February the virus spread throughout the world and scientists, researchers and microbiologists are now saying that we the humans have to live with this virus for some coming years. This virus can enter in our body via nose, mouth or eye. However the main source of spreading the virus is through our respiratory system that involves mouth and nose. Scientists believed that by using of face mask we can decrease the chance of getting infected with the virus by 90%. That is why many countries are now enforcing the law to wear the face mask whenever someone wanted to go out. Some rules are also passed that whoever do not wear the face mask it should be fined. It is very tedious task for a person to keep an eye on every one that if someone is wearing the mask or not. For last few years helped a lot for the human surveillance and security. The efficiency and authenticity of computer vision system has also improved to that extent where humans can rely on. So the detection of humans with or without face mask can became possible through video surveillance cameras. This research propose a system to detect that weather the person in the camera is wearing face mask or not. The system will show the person with face mask with green bounding box and the person without face mask with red bounding box in real time. This application will decrease the human effort on video surveillance and helps in abiding the law to wear face mask.

Problem Statement:

After the corona virus pandemic every country made this rule to wear the mask whenever the person wants to go outside. The surveillance of person with wearing mask will become tedious for humans. This task of surveillance can be done using computer vision techniques. In this research application a system will be introduced that will efficiently detect the persons with or without mask using deep learning based approaches.

Objective Statement:

The main objective of this system is to reduce human effort on video surveillance and made it automatic by efficiently detecting the persons with or without mask.

Proposed Solution:

Efficient of detection of person wearing mask is now became a hot topic in the field of computer vision. This research propose a novel and real time surveillance system to effectively detect the person wearing mask or not.

Summary:

With the advent of corona virus pandemic, the researches proposes to wear mask for the safety because about 90 percent of virus is spread via respiratory system and wearing mask can protect us to not to get effected. So wearing of mask has become the international law and imposing of fine is being done who is not wearing mask. To detect the persons from a crowded area with human surveillance is quite a trivial task. So human vision system can be used for this problem. As now a days computer vision are now helping a lot in video surveillance along with humans. So this research propose an application of detecting the persons with masks.

Approach:

The proposed solution is to made a computer vision system that will effectively identifies that weather the person is wearing the face mask or not. This will be achieved using several computer vision and machine learning based techniques. First of all the human face will be detected using some detection algorithms then some computation will be done on that detected face image. Based on these computations the computer will identify that if the person is wearing mask or not. There will be two phases of this system

  1. Training
    1. Load dataset having person with face mask and without face mask.
    2. Train the face mask classifier using a CNN based model.
    3. Save the trained model for classification.
  2. Application
    1. Input an image or video
    2. In case of video as an input split the video into the corresponding frames.
    3. On each frame detect that if the frame has some human face in it using opencv based haar cascade model.
    4. If the human face is present then classify this region of having face with our classifier.
    5. The classifier will classify along with the bounding box color in real time.

Dependencies and Limitations:

As we know that computer vision systems are heavily dependent on the availability of the data. For the quality application the benchmarked dataset must be available to achieve better accuracy and efficiency. As this problem is quite new so there is a limitation that we did not have much dataset available. That is why the system lacks sometimes. Face should be reasonably visible for the correct detection.

Novelty and Significance:

The system can detect the human with face mask irrespective of gender and age. We can detect the persons with any posture and side.

Technology:

The tools and technology used in this application are Python 3.6

The application can be trained and tested in any core 2 duo or above CPU with 4 GB or above RAM.

Performance Metrics:

Accuracy measure will be used in this application as a performance metric. During training our model achieved 98% accuracy.

Libraries:

The libraries used to develop this application are:

  • Tensorflow 2.1
  • OpenCv 3.4
  • Keras

Dataset Description:

The dataset is having two classes. One is with mask and other is without mask. There are total 1376 images. Out of which 690 are with masks and 686 are without mask. All images are used in training.

Ethical/Privacy Concerns:

This application uses a publically available dataset so ethical or privacy is abided.

Remember, at the center of any academic work, lies clarity and evidence. Should you need further assistance, do look up to our Computer Science Assignment Help

Get It Done! Today

Applicable Time Zone is AEST [Sydney, NSW] (GMT+11)
Not Specific >5000
  • 1,212,718Orders

  • 4.9/5Rating

  • 5,063Experts

Highlights

  • 21 Step Quality Check
  • 2000+ Ph.D Experts
  • Live Expert Sessions
  • Dedicated App
  • Earn while you Learn with us
  • Confidentiality Agreement
  • Money Back Guarantee
  • Customer Feedback

Just Pay for your Assignment

  • Turnitin Report

    $10.00
  • Proofreading and Editing

    $9.00Per Page
  • Consultation with Expert

    $35.00Per Hour
  • Live Session 1-on-1

    $40.00Per 30 min.
  • Quality Check

    $25.00
  • Total

    Free
  • Let's Start

Get
500 Words Free
on your assignment today

Browse across 1 Million Assignment Samples for Free

Explore MASS
Order Now

Request Callback

My Assignment Services- Whatsapp Tap to ChatGet instant assignment help

Hire Certified Experts
Ask your Question
Need Assistance on your
existing assignment order?