APPLICATION OF TEACHABLE MACHINE IN REAL-TIME FACE RECOGNITION

Abstract

These days, AI (Artificial Intelligence) and ML (Machine Learning) are two of
the most promising application areas in the world. However, for a non-IT person, it
is very difficult for them to approach and apply AI or ML in their research. To solve
this problem, the author would like to mention a tool released by Google called
Teachable Machine, Teachable Machine will help us get the most basic experience
of artificial intelligence (AI). Teachable Machines is built based on a convolutional
neural network (CNN - Convolutional neural networks) - one of the advanced Deep
Learning models, it helps us build intelligent systems with high accuracy. Many
users in 201 countries use Teachable Machine to create more than 125,000
classification models [2]. Within the scope of this paper, the author will also build a
machine learning model based on the Teachable Machine tool to build a face
recognition system in real-time. The article will focus on the problem of collecting
data from different objects, then using the Teachable Machine tool to build a
model, and then proceeding to identify people in real-time through the camera.
After 2 times of training the model, the model's accuracy reached 100%.