|
The live wire of Ubona’s service is its state of the art speech recognition platform. The speech recognition platform is completely in-house developed and uses a paradigm different from the prevailing methodologies. At Ubona, speech recognition is performed using our own architecture and algorithms to achieve the following primary goals:
 Reliability of recognition for a very large vocabulary size of hundreds of thousands of words, names and phrases representing shops, establishments, products, campaigns and movie names.
Speaker independence to work in cities of India with millions of people with multitude of languages and accents.
Automation to introduce new shops, establishments, product names, movie names, campaign names, words, phases, etc.
When you call an Ubona service and say “something”, the speech recognition platform first determines what must have been said and converts it into a set of texts. Ubona then prompts for your input on if the recognition is correct and subsequently it performs actions like call, playing the information, sending SMS etc based on your choice. |
|
