Hand Gesture Recognition with Inertial Sensors.

Area of Science:

• * Education
• * Human-Computer Interaction
• * Assistive Technology

Background:

• * Dyscalculia presents significant challenges in fundamental arithmetic skills, particularly with traditional teaching methods.
• * Educational games offer an engaging alternative for learning.
• * Gesture recognition technology, enhanced by wearable sensors, shows potential for improving learning outcomes.

Purpose of the Study:

• * To propose two gesture recognition methods for an interactive educational game.
• * To utilize wearable inertial sensors for hand gesture recognition.
• * To enhance learning for children with dyscalculia through gamification.

Main Methods:

• * Developed a multilayered perceptron classifier using accelerometer and gyroscope data.
• * Created a simplified nearest centroid classifier using only accelerometer data.
• * Integrated gesture recognition into open-source educational games.

Main Results:

• * Both proposed methods achieved over 93% accuracy in recognizing 15 distinct hand gestures.
• * Inertial sensor-based gesture recognition was validated through experimental results.
• * The system demonstrated effective hand gesture recognition for educational game applications.

Conclusions:

• * Gesture recognition via wearable sensors provides an accurate and effective tool for educational games.
• * This approach can significantly benefit children with dyscalculia by offering an engaging learning experience.
• * The developed methods are suitable for low-cost wearable devices, promoting accessibility.