Real time Bangla Digit Recognition through Hand Gestures on Air Using Deep Learning and OpenCV

Abstract

Digit Recognition in real time through hand gestures has achieved great attention in machine learning and computer vision applications. This article focuses on identifying Bangla numerals in the air using hand motions. This research leads to the stairwell, allowing for more investigation in the same subject for various Bangla characters and even phrases. The major issue, however, is coping with the wide range of handwriting styles employed by various users. Many studies have been done on the identification of Bangla handwritten digits, but none has proven successful at recognizing Bangla digits in real time using hand gestures in the air. As a result, this article describes the creation of a Bangla digit recognition model that employs a Convolution Neural Network (CNN) to predict Bangla digits by observing hand movements in the air space. After a thorough examination, the suggested system attained a 98.37% accuracy on the BanglaLekha-Isolated dataset.