About Felipe Clemente
Felipe Costa Clemente is a computational linguist and phonetician who has been dealing with the development of linguistics features for automatic speech recognition (ASR) and text-to-speech (TTS) systems. Although, he has a vast research about phonetics features of R-sounds on Brazilian Portuguese.
Tools
Languages
Skills
General
My Research
Over the last few years (2015), I have been working on a personal project: build my own acoustic and textual corpora aimed to feed ASR systems.
There is a plenty of narrators, representing the great variety of Brazilian Portuguese dialects. These data is manually transcribed and evaluated by automatic tools which garantees its qualities. Furthermore, I will be recording different speakers, all around the country. The objective is to balance this corpora with different genders, age and locations.
This corpora will have millions of text lines too, all data from open sources textual corpora and they will be prepared to generate n-grams language model. The main objective is to cover different text styles and provide a generous model to recognize Brazilian Portuguese text patterns.