Separate application DROTR-Local envisaging that the User records a sample of his/her voice only on one occasion and sets his/her language in the profile. In the same manner, he/she records a voice sample and sets the language (on one occasion only) for another person, with whom he/she intends to establish contact and have a conversation. After that, DROTR-Local application is launched, the device is put between the two persons and their conversation is translated without the need for any other actions; DROTR-Local takes part in the conversation actually as a human being - interpreter – in fact, as the third «individual». The speaker identification algorithm analyses every uttered phase and defines who of the speakers said it. No need to press the buttons or wait until you can or can not speak; the recognizer always listens to you, and by means of the speaker identification feature we determine the language which should be recognized. Besides, in a short while we will use the speaker identification in the following areas: automatic stenographing of interviews and conferences (the algorithm is capable to identify stably up to 5 speakers).



© Pavel Naumenko / Library and Information Science Resources 2014