Hand Gesture Recognition based on Electrical Impedance Tomography Measurements using Genetic Algorithms

Electrical Impedance Tomography (EIT) for gesture recognition is on the rise in the bio-medical field. It is an imaging technique based on a set of electrodes which are placed on the forearm of the individual. The inner conductivity distribution inside the forearm will be calculated based on measured boundary voltages resulting of the injected excitation signal. In this work, the Genetic Algorithm, using the Non-Blind Search has been developed as a solution for the EIT Inverse Problem. Therefore, several measurements have been conducted for American sign numbers (0-9) for gesture recognition applications. The reconstructed images were used in a Convolutional Neural Network (CNN) classification algorithm for gesture recognition. The overall accuracy achieved for image recognition was 93%.