Mask Detection

Table of Contents

1. Problem Statement
2. Description
3. Technology used
4. Contributors

Problem Statement

During these unprecedented times, for the sake of safety of everyone, ensuring the people wear masks at public places becomes utmost critical and there should be a simple mechanism for it. Here, with the help of powerful Machine Learning field, we suggest some models which could detect whether a person is wearing mask or not.

Description

• Pre-Processing

  All Images were converted to 120 X 100 size and Dimensionality Reduction techniques like PCA and LDA were applied to the them

• Training and Testing

  3 different classifiers were used to train the dataset and test it on testing dataset . Classifiers used were : 1)SVC 2)MLP 3)KNN

• Comparision

  Among the classifiers used, SVC classifier was the best followed by MLP classifier and KNN classifier respectively.

• Prediction on an unseen image

  On receiving an unseen image, firstly its face area is extracted with the help of Mediapipe framework , followed by feature extraction by PCA/LDA , and then finally feed to the classifiers, which finally detects the mask in the image if present

Technology Used

• Python
• Sklearn
• Numpy
• Pandas
• Mediapipe Framework

Contributors

• Patel Darsh Babubhai (B19CSE115)
• Savani Hard Hareshkumar (B19CSE080)