Improving Real-time Hand Gesture Recognition System for Translation: Sensor Development
Sohal Islam, Showni Rudra Titli, Kazi Arham Kabir, Md Abdullah Al Hossain, Md. Altaf Hossain · 2022
A lot of deaf-mute persons rely on sign language as their primary form of communication, which can be difficult for the majority of us to comprehend. The Deaf community’s standard language of communication is currently Sign Language (SL), which is articulated via the use of body motions (hand, face, torso) as their main form of communication and observed through the eyes. The hands convey the majority of the information in sign language.Automated sign language recognition systems that use vision-based technology must extract crucial hand features from real-time image sequences in order to provide comprehensive and accurate categorization. A gesture recognition glove is proposed for real-time translation of English International Sign Language in our proposed method. Firstly, the smart glove will frame-by-frame watch the sign movements. The acquired data will then be analysed and the necessary linguistic data extracted. Secondly, to meet the desired purpose for deaf and mute people words are displayed through an LCD along with voice output.The method includes a gesture extraction phase that uses two types of sensors. The hand gesture image is processed in the approach to distinguish distinct characters, and we construct our own structural circuit design. However, a smart glove prototype graphical user interface application efficiency was tested based on the real-time data pre-processing and output visualization. Finally, both approaches are subjected to the gesture recognition step, the lowest efficiency was 92 % and the highest efficiency was 100%. The primary goal of this study is to introduce a device to society and encourage people with disability to communicate and explore the real world.