Malayalam Sign Language Recognision
Communication holds a central position in social interactions. In a predominantly aural society, sign language users are often deprived of effective communication. The National Institute of Speech and Hearing(NISH) recently introduced Signs for Malayalam Alphabets. Applications for recognising previously formed sign languages are already available. Here we aim to develop an assisting system that recognizes hand gestures in Malayalam and converts them to textual format. Several solutions have come up but none of them have been portable for them to be used in a standalone device or application. We have alleviated this problem by harnessing the power of the mobile phone and the recent advances in deep learning. Our work emphasizes Dataset creation for this newly formed sign language, training a model using squeezenet to recognize these signs, and integrating the model into a flutter application for ease of access.
Teachers and students at NISH, under the supervision of sign language experts there, have come up with the finger-spelling, which covers vowels and consonants too. This new sign language alphabet in Malayalam will eliminate the challenge of teachers in conveying messages through lip movement. This would also help bring about a qualitative change in the lives of those who are hearing impaired.
This project aims at introducing a mobile application that recognizes these Malayalam sign alphabets with the help of machine learning technology. Since these Malayalam Sign Alphabets are new, there was no proper dataset for this system. So the creation of a dataset for Malayalam Sign Alphabets is also an objective of our project.
Codebase for Sahayi application