Senyales: FSL Translator for Day-to-Day Emergencies
Proceedings of 2021 The 11th International Workshop on Computer Science and Engineering
(2021), pp. 255-264
Alain Vincent Mindaña
a
,
Daian Villa
a
,
Marie Sheila Martin
a
,
Qaeda Vennett C. Cinco
a
,
Anthony D. Aquino
a
,
Shaneth C. Ambat
a
a College of Computer Studies, Far Eastern University Institute of Technology, Manila, Philippines
Abstract: In the Philippines, there is still a barrier language between the deaf and hard of hearing and abled communities. This could be attributed to how time and resource-consuming it is to learn FSL. This study aims to develop an FSL emergency sign detection in hopes to diminish the prejudice against dhh and the barrier that prevent them from being entirely accepted by non-dhh. To test the hypothesis of how fast hand detection works in emergencies using mobile, YOLO v4 tiny was used as the main algorithm for FSL detection for selected phrases often used in emergencies backed by an online survey for its usefulness. The results yielded an average accuracy of 85% in detecting the hand signs in different background environments with ample lightning and that using the mobile application to detect emergency hand signs is fast and accurate enough when in different backgrounds.