Publication: Deep convolutional neural network approach towards a smart home automation system
Loading...
Date
2022-08-01
Authors
Zar’ai, Muhammad Fadhil Iman
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
A home automation system regulates the temperature, appliances, entertainment systems, and lights. The Internet of Things comprises gadgets and sensors linked together by a common infrastructure. A home automation system connects multiple controlled devices to a centralized server. A tablet or smartphone application can be used to access the user interface for controlling and monitoring these devices, which can also be accessed from a distance. With the theme of Harmony Home, this article proposes an architecture that serves as a proof-of-concept for emotion recognition and regulation in smart home environments. The project aims to identify a person's emotional state using facial expression analysis. The system then offers a design that is intended to control these emotions and, where feasible, shift them toward a positive mood. The state-of-the-art in emotion control through music and colour/light is used to improve human care and quality of life. A temperature and humidity control feature is also added to produce a comfortable and cool environment. Five facial emotions need to be identified: fear, anger, happiness, neutral, and sadness. The model was trained to recognize these emotions using the DCNN algorithm. As a result, a final model accuracy score was 69 percent, with Neutral emotion receiving the highest score.