INDIAN SIGN LANGUAGE DETECTION USING ARTIFICIAL INTELLIGENCE

Abstract

In the realm of human computer interaction, sign language detection has emerged as one of the most significant study fields. Deep learning and machine learning can also be utilized to help with this problem. This may be quite useful in talking with others for the deaf and dumb. Physically challenged persons can convey their thoughts and emotions through sign language. The Indian sign language dataset has been subjected to eight machine learning approaches: SVM, Random forest, KNN, Decision tree, Logistic regression, XgBoost, LightGBM, Naïve byas and one deep learning approaches: CNN. The dataset is also subjected to pre-trained model ‘ssd_mobilnet_v2_fpnlite_320x320_coco17_tpu-8’ for real time detection of Indian sign language. Furthermore, the study intends to use comparative analysis of performance to achieve its goals to make the selection of relevant prediction techniques easier.