Holidays are a great opportunity to make ideas which go on in the mind come true. It was a month that I was thinking on a simple practical implementation of ML techniques together in one application. Last holiday, I put some effort to make a tiny bot in Python that I named it WYNA, stands for “What’s Your Name App”. The goal was to make use of both speech & face recognition tools in a single app in order to have integrated basic interactions with human. All it does is that, it finds nearest person among people in the screen, asks the person’s name, then, after the person says his/her name, WYNA replies with a random greeting, and remembers the person’s face and name in the next encounters. Undoubtedly it has long way to be called a bot, but for me it was as amazing as Boston Dynamics robots 😅.
For face recognition I used FaceNet model in Keras, and for speech recognition a Python package named SpeechRecognition version 3.8.1 is used which has several online and offline methods for recognizing speech including CMU Sphinx , Google Speech Recognition and Microsoft Bing Voice Recognition.