Gesture Recognition in Indian Sign Language Using Deep Learning Approach
Perkampally Sanjana Reddy, N Sai Kumar, Bonthagarla Manikanta Sai Teja, Andraju Bhanu Prasad, Shanmugasundaram Hariharan, Vinay Kekreja · 2024
This research study intends to determine how deep learning techniques can be used to detect Indian sign language in real time. Indian sign language (ISL), a unique visual gesture language, is used by deaf and hard-of-hearing persons in India to communicate with one other and with people who can comprehend sign language. ISL plays a major role in communication not only among deaf people but among deaf and normal people. We take the dataset images from the webcam using the OpenCV and Media Pipe which not only helps us to create but to locate the key points of the image data. The recognition is easier by using deep learning methods. In this work, deep learning algorithms such as training an LSTM DL model and making a real-time prediction using sequences and output in the format of text/visual. The evaluation is done using the Confusion Matrix and accuracy. It has been found that the accuracy of this model is 91 percent.