A Review on Sign Language Recognition Systems

Providing the software which aims to bridge the gap between the inarticulate community and the rest of the world for easing the communication is what is the prioritized goal of Sign Language Recognition software. Open CV for hand recognition, Gaussian blur, and contour approximation technique is used for the enhancement of the recognized gesture. For the accurate prediction of the software, the TensorFlow framework is best suited with a CNN model of VGG-16 architecture and the training method of Transfer Learning. Usage of these tools leads to the accurate prediction of sign language gestures into the English alphabet with higher accuracy. The technology stack is robust and scalable, considering the future scope of the project. Various methods and structures for sign language recognition are discussed in this review paper. This will aid researchers in identifying the most suited approaches or algorithms for their application.