Mobile Voice Recognition

Mobile devices have some annoying limitations. In particular, data input using tiny keyboards is cumbersome. Output capabilities and screen size are less of a problem. If you could only talk to the device and get results displayed on the screen you could make phones much more user friendly.

The topic of my master thesis was speech recognition using neural networks. It was a while ago - I will not mention the exact number of years :-). Technology has made a lot of improvements since then. Many companies have deployed speech recognition in call centers to replace traditional IVR systems and improve customer experience. Desktop products such as Dragon Naturally Speaking have been available for a number of years. Underlying standards such as VoiceXML have matured.

Speech recognition is also making progress in the mobile space. Nuance, the leading provider of speech solutions, is promoting its Mobile Speech Platform. Applications include mobile search and mobile messaging. Search responds to voice commands such as

“Find restaurant near Union Square, San Francisco”
“Weather in the Bay Area”
“Call Bob Tekiela”

Messaging combines dictation software with Text To Speech (TTS) engine enabling users to speak and/or listen to emails, IM or text messages.

Other companies working on mobile speech recognition include:
Promptu
V-Enable
VoiceBox Technologies
VoiceSignal