Speech to text Community Application using Natural Language Processing
Rakesh Kumar, Meenu Gupta, Sandalika Sapra · 2021 5th International Conference on Information Systems and Computer Networks (ISCON) · 2021
Android-based mobile applications are on-demand these days. People in a community can communicate with each other through internet-based platforms using cloud firebase. Since typing messages is a cumbersome activity, a speech to text application would cater for a hassle-free mode of communication. A flutter based android application integrated with natural language processing can capture the spoken words and digitize them into textual format. The efficacy of the application depends on the vocal parameters of the speaker that includes- pitch, loudness, intonation etc. On the basis of the audio captured, the application processes the text and displays it on the screen. The application uses a cloud firebase for user authentication and sign-in. Once a user registers on this application, they can communicate with others using the speech to text feature. The performance of the application is evaluated using hidden Markov models that take variegated forms of audio input into consideration. The precision of the application is 91.5%, recall is 95.4%, f1 score is 86.8% and accuracy is 89%.