ABSTRACT
In air writing, hand gestures in three-dimensional space are converted into visual linguistic characters. This technique has already taken a place as an intuitive, simple, and natural way of Human-Computer Interaction (HCI). Numerous researches on air writing have been done where various algorithms are proposed. However, most of the researches are based on English character recognition by air-writing. To the best of our knowledge, very few researches are done on air- writing for Bengali characters. This research represents a hand gesture-based real-time classification model for Bengali character recognition in air-writing. A data acquisition device is developed to collect real-time data of hand movement. This device reads the orientation and acceleration of hand gestures in three-dimensional space when a user writes a character in the air. These data are transmitted to a computer in real-time via Bluetooth. A data set is prepared taking the real-time air writing data of Bengali characters. This data set is used to train a K-NN classifier. The accuracy of the classification model is 96.7%. Finally, the trained classification model is used in real-time air-writing recognition of Bengali characters.
