Open AccessFacemask Detection using Inception V3 Model and Effect on Accuracy of Data Preprocessing Methods
Author(s) -
Yongyuan Li
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/2010/1/012052
Subject(s) - preprocessor , computer science , artificial intelligence , detector , data pre processing , data mining , pattern recognition (psychology) , image (mathematics) , machine learning , telecommunications
Nowadays, image classification done by Machine Learning can classify images within an instant after an efficient model is built. Such techniques can help identify whether a person correctly puts a mask on. During the Covid situation, it is important to ensure the people in public areas put on a mask correctly so it can cut off the route of mass infection. In this paper, three classes of image classification were tested: with mask, without mask, and mask-weared-incorrectly. Based on the Google Colaboratory platform and Inception V3 model, a three-classes-detect-model with 94.52% testing accuracy was built. In addition to building the models, the effectiveness of data preprocessing has also been tested. After applying different methods for data preprocessing, the testing accuracy for the two-classes-detector model improved to 97.11% but the three-classes-detector model decreased to 94.26%.