Sign language is a silent and visual language that requires us to use gestures and facial expressions that enable us to communicate with individuals who do not have the ability to hear or speak. Each country has its own spoken language-specific sign language. In this study, a data set consisting of 2974 images, including 29 Turkish sign language alphabet characters, was used. Sign language is not known by all segments of society. This situation causes communication problems in the social environments of hearing-impaired people. An individual who is not deaf but does not know sign language has the same problem. Thanks to deep learning technologies, object recognition and classification studies have gained speed day by day as a result of the analysis of these objects. The aim of this study is to classify Turkish sign language alphabet characters using deep learning models. For this purpose, the identification of sign language alphabet characters with CNN and Inception V3 was carried out in this study. According to the performance criteria, the accuracy rate was 76.20% in the CNN model and 97.50% in the Inception V3 model. As a result, with this study, it was seen that deep learning models were successful in classifying the Turkish sign language alphabet.
Anahtar Kelimeler: Deep learning, Convolutional neural networks, Turkish sign language
|
