Deep Learning Method for Sign Language Recognition: A Systematic Literature Review

This systematic literature review will discuss and explore the implementation of different types of artificial intelligence models including deep learning in recognizing and interpreting sign language. Nowadays, communication is an important aspect of human life. However, the communication ability of each human being is not the same, some have limitations in communicating such as deaf and speech impaired people so they need tools such as sign language. Unfortunately, not everyone can understand sign language and it has become a limitation in communication. This encourages researchers to continue developing various systems to solve this problem. In this review, we conducted an in-depth evaluation of 14 published articles that have been published with a focus on developing artificial intelligence systems for sign language recognition. The review was conducted by looking at the models, techniques, accuracy, and datasets used in each article. The expected results will be used as considerations for the creation of more advanced systems in the future. The results show the challenges of AI in understanding complex sign language gestures and expressions, but also show positive progress with various AI models, where ResNet-50 shows the highest accuracy rate reaching 99.98%.