Hand Gesture Recognition using Electromyography Signals (EMG) and Artificial Intelligence
Field: Artificial Intelligence
Abstract: This research project pursues developing general and user-specific models for hand gesture recognition using EMG signals measured on the forearm muscles. For model development, we use artificial intelligence techniques. In addition, the project proposes to develop a protocol for the evaluation of the accuracy gesture recognition models including a protocol for data acquisition.
Project code: CEPRA XIII-2019-13
Funder: Corporación Ecuatoriana para el Desarrollo de la Investigación y la Academia (CEDIA).
Partners: Escuela Politécnica Nacional (EPN), Universidad Técnica de Ambato (UTA), Universidad de las Fuerzas Armadas – ESPE.
Start date: August 15, 2019
End date: November 15, 2020
Director: Dr. Marco E. Benalcázar
Hand Gesture Recognition using Electromyographic Signals (EMG) and Artificial Intelligence and its application to the implementation of human-machine and human-human interfaces
Field: Artificial Intelligence and Machine Learning
Abstract: In this project we propose developing recognition systems for 11 static and dynamic hand gestures. The proposed system will have as input both electromyographic signals (EMG) measured on the forearm muscles and orientation signals measured on the forearm using an inertial measurement unit (IMU). The development of the recognition algorithms will be based on the use of machine learning techniques. Finally, as an example which demonstrates the applicability of the proposed systems, we will develop (1) a human-machine interface for video game control through gestures, and (2) a human-human interface that allows a user to control the movement of another user's hand.
Project code: PIGR-19-07
Funder: Escuela Politécnica Nacional (EPN)
Partners: Departamento de Informática y Ciencias de la Computación (DICC), Departamento de Automatización y Control Industrial (DACI), Departamento de Electrónica, Telecomunicaciones y Redes de la Información (DETRI)
Start date: May 04, 2020
End date: May 04, 2022
Director: Dr. Marco E. Benalcázar
Classification of Electromyographic Signals of the Human Arm using Pattern Recognition and Machine Learning
Field: Artificial Intelligence and Machine Learning
Abstract: The muscular activity of the human arm produces electrical signals called electromyographic signals, or simply electromyography (EMG). The classification of these signals has multiple application domains including interfaces for video games, robotics, sign language translators to text or speech, bionics, among others. Several works have been proposed in the literature for the classification of these signals. The main problems of these proposals are the following: (i) limited number of gestures that these systems can recognize and, (ii) the overestimation of their classification accuracy caused using few test samples. In this project, we propose developing new EMG classification models using pattern recognition and machine learning techniques. For this project, EMGs will be acquired using the commercial Myo armband. The proposed models will be able to learn to classify a person's gestures through machine learning techniques. This will allow the system to be adaptable to the diversity of EMGs for the same gesture between different people.
Project code: PIJ-16-13
Funder: Escuela Politécnica Nacional (EPN).
Partners: Departamento de Informática y Ciencias de la Computación (DICC).
Start date: April 01, 2017
End date: January 31, 2020
Director: Dr. Marco E. Benalcázar