Desenvolvimento de um protótipo de sistema web para elaboração de laudos médicos utilizando sistemas de reconhecimento automático de fala

Abstract

The overall purpose of automatic speech recognition systems is to allow the interaction between humans and electronic devices through speech, for example, the content captured from user's speech using a microphone, can be converted into textual transcription. In general, such systems should be able to overcome adversities such as noise influence, communication channel variability, age, accent and speed of speech, concurrent speech from other speakers and spontaneous speech. Even with this challenging scenario, this work aims to investigate this technology to be employed in the medical field. The research was carried out through a systematic review for the scientific basis of concepts and identification of relevant research studies. To validate the feasibility of using these systems in the medical field, a Web System Prototype was developed to generate medical reports through automatic speech recognition for the Brazilian Portuguese Language. The prototype was developed with the application of Delivery in Stage technique, from Software Engineering. The selection of the automatic speech recognition technology to be integrated to the prototype was based on an evaluation of the accuracy rate of seven systems for the Brazilian Portuguese Language. The development of this work allowed to conclude that these systems of automatic speech recognition can be used in the medical environment, providing support not only for making medical reports, but also acting as a monitoring record during a medical procedure.