Publication

Abstract : For people with physical disabilities, especially those who are confined to a wheelchair, the development of human-computer interaction has created new opportunities. Due to their restricted mobility, these individuals continuously face significant challenges in accessing their everyday home appliances. When we consider the conventional systems based on switches, they require physical assistance, which makes it difficult for these people to use them. In this paper, the development of a real-time intelligent home automation system using hand gestures is presented. This work uses a simple Convolutional Neural Network (CNN) made using Tensorflow and Keras libraries for gesture recognition, which achieved a testing accuracy of 99.5%. Further, this ML model was deployed in an ESP32 CAM module for image acquisition to detect the gesture given by the user. The gestures that were recognized helped in operating specific home appliances, enabling touchless control and thereby improving the quality of life of people with physical disability.