Sign Language Recognition Systmem
This project aims at identifying alphabets in Indian Sign Language from the corresponding gestures.
Created on 27th June 2021
•
Sign Language Recognition Systmem
This project aims at identifying alphabets in Indian Sign Language from the corresponding gestures.
The problem Sign Language Recognition Systmem solves
A dataset with more variation and a higher quality can really boost the accuracy of our current models. Also we think that using more complex models like artificial neural networks,
or applying deep learning on the HOG vectors should improve the accuracy as they are capable of extracting richer information from these vectors. Increasing the levels of heirarchy
with proper hierarchy levels created on the basis of which nodes are being misclassified can
result in improvement in accuracy, but we can not surely claim that as suggested by multiplicative probability rule in our experiment with hierarchical classification.
Challenges I ran into
The Indian Sign Language lags behind its American Counterpart as the research in this
field is hampered by the lack of standard datasets. Unlike American Sign Language, it uses
both hands for making gestures which leads to occlusion of features. ISL is also subject to
variance in locality and the existence of multiple signs for the same character. Also some
character share the same alphabet(E.g V and 2 have the same sign, similarly W and 3 have
the same sign) and the resolution of the sign is context dependent.
Technologies used