Development of a Face Mask Type Detection with Multiclass Classification using Artificial Intelligence on Python Anindya Ananada Hapsari, Halimatuz Zuhriah, Devan Junesco Vresdian, Onki Alexander, Brainvendra Widi Dionova, Untung Suprihadi
Universitas Global Jakarta
Abstract
This research presents a system based on a machine learning system with artificial intelligence. The approach presented uses a camera to detect people who are not wearing face masks and it is hoped that it can detect type from mask. The purpose of this study is to be able to compare whether a computer system can correctly distinguish different types and types of masks using multiclass classification with artificial intelligence and training datasets on the system. And testing with different amounts of data does it affect the detection sensitivity of the system. The main contributions of this research are creating a prototype face mask detector which is implemented in three phases to help detect the presence of a face mask detector in real-time using images and video streams. The data used is a dataset consisting of images using two dataset contain two categories and four categories type of mask. This study also tried to conduct training on a dataset of several different types of masks. The model created has used deep-learning methods to develop classifiers and collect photos of individuals wearing masks and not wearing masks. Mask and non-mask class. Which is then implemented in Python with Google Colab environment along with the Open-CV modules, Keras, NumPy, tensor flow, sci-py, and matplotlib. Trials will be carried out with the dataset training process first using CNN and then collecting the data system accuracy.
Keywords: Artificial Intelligence- CNN- Face Mask Detection- Multiclass Classification- Python
