AIR HANDWRITING BY USING CNN MODEL

Abstract

Gesture recognition has gained significant attention due to the growth of IoT and smart device technologies. Among its various challenges, air-writing—writing characters in mid-air using hand movements—stands out as a complex yet important task. This paper presents a wearable air-writing system that allows users to write English letters freely in three-dimensional space, without requiring strict adherence to writing conventions. The system is based on data from an Inertial Measurement Unit (IMU) and uses Dynamic Time Warping (DTW) as the core algorithm for gesture recognition. To improve accuracy and optimize the combination of IMU data with DTW, we propose a novel adjustment method that enhances system performance. Experimental evaluations show that the system achieves a recognition accuracy of 84.6%. Moreover, the results reveal that the effectiveness of DTW-based recognition varies between users, highlighting the importance of adapting the system to individual usage patterns.