Abstract：Objective To design an intelligent “eye-speech” assistive device for aphasic patients, which can output speech phrases in
a variety of communication situations controlled by the direction of the patients’ individual eye movements, allowing the patient to
achieve basic communication and meet real-life needs. Methods Based on the Arduino platform, recognition of right eye movement
states (up, down, left, right and closed) was achieved using four infrared reflectance sensors connected to regular glasses. For each
combination of two consecutive eye movements (up, down, left, and right), the text-to-speech module (Parallax Emic 2) outputs the
predefined phrase corresponding to the combination. Results The power supply voltage test result of key components was +4.7~4.8 V, and
the power supply was normal. All sensor outputs, USB pins and Arduino pins connection were turning on. The best sensing distance
of infrared external reflectance sensor for the direction of single eye movement was 3~4 mm, and the maximum sensing distance
was 9~10 mm. 128 predefined phrases corresponding to the combination of the 2 directions of right eye movements in the main
control function mode and the scenario mode, the normal rate of system speech output was greater than 93% in the 60 times of test.
Conclusion The “eye-speech” solution and device designed and made by combining the eye movement direction capture technology
with the text-to-speech technology has wide range of applications and good expansibility.
雷涛a，许兆坤b，王昭昳b，李钊a，张林媛a，路国华a. 用于失语群体的智能“眼-语”助手的设计[J]. 中国医疗设备, 2022, 37(7): 35-38.
LEI Taoa, XU Zhaokunb, WANG Zhaoyib, LI Zhaoa, ZHANG Linyuana, LU Guohuaa. Design of an Intelligent “Eye-to-Speech” Assistant for Aphasic Groups. China Medical Devices, 2022, 37(7): 35-38.
