Multi language voice commanding and Interactive Slides


A little bit more voice commanding with picoFlamingo, this time using CMU PocketSphinx and featuring multi-language support, that is, dynamically switching between different language at run-time. This feature makes picoFlamingo the first presentation solution for international organisations as UN or EU... just kidding :). A video demonstration after the break

The demonstration is running in a x86 laptop. We are having some logistics problems to use our BeagleBoards and OMAP ZoomII lately, but hopefully a demonstration of full-fledged picoFlamingo will be available soon.

For the voice recognition we just made a simple modifications to the pocketsphinx_continuous example distributed with pocketSphinx. Then the recognised words are piped to a filter (a simple C application) that transform them into commands picoFlamingo can understand. This filter program also takes care of the language switching and the deactivation of the voice recognition.

If you speak Spanish or Galician and want to play with speech recognition here are some additional information. For Spanish we are using the files in this page. Galician acoustic model is pretty similar and choosing carefully the commands we want to use, the Spanish data will work without problems. We just needed to update the dictionary in that case.

The picoFlamingo Team