Why Type When You Can Speak?
If you have utilized Google Now or Apple’s Siri, you know how effortless it is to say something to the handset and view it being executed. Picture if you might do this in Tamil or Bangla? The rising effect of AI (artificial intelligence) on the correctness of voice and speech recognition and the elevated need for multi factor authentication is boosting companies to aim on regional languages. Google last week declared extension of its capabilities for speech recognition to support more than 30 fresh languages, comprising 8 languages form India such as Bengali and Kannada. Indus OS also lately declared speech to text abilities for more than 23 languages from India. With companies aiming on the upcoming billion first time handset consumers, they are gambling on voice.
Liv.ai, the Bengaluru based company, provides an app on Android platform for conversion of speech to text in languages comprising English, Hindi, Punjabi, Bangla, Telugu, Kannada, Malayalam, and Tamil that will be rolled out next week. With 2 additional devices scheduled for roll out by the end of this year, the company is seeing elevating interest. “Our servers were congested when we rolled out the app that established the demand in the market. The cognitive burden related with typing is boosting users to take to speech,” claimed cofounder and CEO of Liv.ai., Subodh Kumar, to the media in an interview.
With more than a million hits each day, Liv.ai is seeking to bring that to a billion by end of next year. Hindi, English, Kannada, and Marathi are indicating capable growth presently.
Speech recognition has arrived a long way from recognizing a handful of fundamental words such as text, call, and so on to being crucial in apps that are used for conversion of speech to text. “Originally, when we rolled out speech recognition, it consisted 95% accuracy for fundamental 10–15 words and presently there is 80% accuracy even with 1000 words. This is due to AI and machine learning making the systems enhanced,” claimed co-founder of Reverie, Vivek Pani, to the media in an interview.
The voice and speech recognition market in 2017 is anticipated to be esteemed at $6.19 Billion and is probable to cross $18.30 Billion by the end of 2023, at a rate of 19.80%. Conversion of speech to text also makes way for transcription apps such as Brainasoft that operates in international languages comprising German and Spanish. The apps can also cancel out noise in background.