WINAS is a system implementation for over-the-air handwriting recognition. The system translates commercial Wi-Fi APs to radars with passive sensing capability. It utilizes hand reflected Wi-Fi signal for reconstructing the hand moving trajectory, from which the system further generates handwriting images. It lastly utilizes existing image classification techniques for handwriting recognition.
The current implementation is capable of supporing multiple languages including handwriting digits, English alphabets and Chinese characters. The performance is robust when experimenting with users at different ages and genders (so with different handsize), and at different environments.
WiRITE: General and Practical Wi-Fi Based Hand-Writing Recognition
@ARTICLE{10098632,
author={Zhang, Yanbo and Sun, Weiping and Li, Mo},
journal={IEEE Transactions on Mobile Computing},
title={WiRITE: General and Practical Wi-Fi Based Hand-Writing Recognition},
year={2024},
volume={23},
number={4},
pages={2943-2957},
keywords={Training;Wireless fidelity;Trajectory;Databases;Image classification;Adaptation models;Mobile computing;Antenna measurement;CSI sanitization;gesture recognition;human computer interaction;wi-fi sensing},
doi={10.1109/TMC.2023.3265988}
}