Recognizing Bengali Sign Language Gestures for Digits in Real Time Using Convolutional Neural Network

Abstract

Abstract—Recognizing sign language gestures for different<br> languages has been found as a promising field of research that<br> explores the possibility of communication by interpreting various<br> signs and translating them into text or speech. Establishing a<br> better communication way between deaf-mute people and ordinary<br> people is the prime objective of this research arena. There<br> are many existing Sign Language Recognition (SLR) systems<br> throughout the world and these SLR systems are implemented<br> using various methods, tools and techniques with a view to<br> achieving better recognition accuracy. This research work aims<br> at applying the concept of Convolutional Neural Network (CNN)<br> for recognizing Bengali Sign Language gesture images for digits<br> only in real time. Bengali sign language images for digits are<br> collected from different individuals and the CNN model is trained<br> with these images after performing several pre-processing tasks<br> i.e. resizing to a specific dimension, converting these RGB<br> images to the gray scale images, finding the equivalent binary<br> images and rotating the images into different degrees both in<br> left and right direction. The experiment is conducted using<br> two major techniques. Firstly, the model has been trained with<br> the dataset containing the equivalent binary images of the row<br> images collected directly from different individuals. Secondly,<br> the dataset is enriched by rotating all the images into 3°, 6°,<br> 9°, 12°and 15°in both left and right directions. After applying<br> the rotation technique, the recognition accuracy is found to be<br> increased significantly. The maximum recognition accuracy of<br> the proposed CNN model with the dataset without image rotation<br> technique is 94.17% whereas the recognition accuracy is 99.75%<br> while including the rotated images in the dataset.<br> Index Terms—Bengali Sign Language Recognition, Convolutional<br> Neural Network, Image Recognition, Real Time Recognition