A Mediapipe-Based Hand Gesture Recognition Home Automation System

As one gets older, his/her mobility tends to decrease. Therefore, simple tasks such as getting up to switch the lights on or turning the fan off can become difficult. Thus, it became imperative to create a system which allows them to perform these tasks - a “Hand Recognition based Home Automation System”. Starting with the research unraveled different ways of implementation of the hand gesture recognition. After considering the functionalities and drawbacks of various methodologies, a library called MediaPipe, was the one that resonated best with this project. This paper includes analysis and comparison of various types of models on a hand gesture dataset - HaGRID.