Naslov
Off-line Inertial-Sensor Based Hand Gesture Recognition and Evaluation

Autori
Kundid Vasić, Mirela ; Grujić, Tamara ; Stančić, Ivo ; Musić, Josip ; Bonković, Mirjana

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Complex Control Systems Vol. 4, No. 1, 2022, 1-6 / - Sofija, 2022, 1-6

Skup
International Conference on Statistics and Machine Learning in Electronics (ICSMLE 2022)

Mjesto i datum
Sofija, Bugarska, 12.05.2022. - 13.05.2022

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
hand gestures, inertial sensors, machine learning algorithms, off-line classification, evaluation

Sažetak
Gesture recognition is a topic in computer science and language technology with the goal of interpreting human gestures with computer programs and many different algorithms. Any bodily motion or state that most commonly originates from the face and/or hand can be interpreted as a human gesture. Most of the research today focuses on emotion detection and recognition of hand gestures using cameras and computer vision algorithms. Gesture recognition can be seen as the way computers begin to understand human body language. There are many different areas this topic of computer science can be applied to ; the main field is human-computer interaction interfaces (HCI). Today the main interaction tools between computers and humans are still keyboard and mouse. Gesture recognition can be used as a tool for communication with the machine and interact without any mechanical device such as keyboard or mouse. In this paper, we present the results of a comparison of five different machine learning classifiers in the task of human hand gestures recognition. Gestures were recorded by using inertial sensors, gyroscopes, and accelerometers placed at the wrist and index finger. One thousand and eight hundred (1800) hand gestures were recorded and labelled. Six important features were defined, for the identification of nine different hand gestures, using five different machine learning classifiers: Logistic Regression, Random Forests, Support Vector Machine (SVM) with linear kernel, Naïve Bayes classifier, and Stochastic Gradient Descent.

Izvorni jezik
Engleski

Znanstvena područja
Elektrotehnika, Računarstvo

POVEZANOST RADA