Continuous sign recognition of brazilian sign language in a healthcare setting

Abstract

Communication is the basis of human society. The majority of people communicate using spoken language in oral or written form. However, sign language is the primary mode of communication for deaf people. In general, understanding spoken information is a major challenge for the deaf and hard of hearing. Access to basic information and essential services is challenging for these individuals. For example, without translation support, carrying out simple tasks in a healthcare center such as asking for guidance or consulting with a doctor, can be hopelessly difficult. Computer-based sign language recognition technologies offer an alternative to mitigate the communication barrier faced by the deaf and hard of hearing. Despite much effort, research in this field is still in its infancy and automatic recognition of continuous signing remains a major challenge. This paper presents an ongoing research project designed to recognize continuous signing of Brazilian Sign Language (Libras) in healthcare settings. Health emergency situations and dialogues inspire the vocabulary of the signs and sentences we are using to contribute to the field