[INQ. NO. 2406E01] A voice-recognition AI real-time caption solution appeared at the 32nd Korea International Broadcasting, Media, Audio & Lighting Show (KOBA 2024), recently held at COEX in Seoul.
This solution introduced by TINNO Co., Ltd. combines an existing streaming engine with voice-recognition AI software, developed in collaboration with an artificial intelligence (AI) company to generate and transmit caption.
TINNO is a company specializing in real-time Internet broadcasting and conference broadcasting, and this solution was also developed targeting the decision-making process that takes place in conferences.
Currently, only real-time Korean caption can be transmitted, but the company plans to allow caption to be converted to a ‘.txt file’ after the meeting ends.
An important task is also indicating who the speaker is. TINNO is considering a wide range of methods, including designating speakers for each microphone, and recognizing speakers through AI facial recognition.
A spokesperson for the company explained, “We are continuing to learn voice recognition, and the accuracy is over 95%,” adding, “In the future, we plan to add a real-time foreign language translation function, or a function that recognizes a dialect and converts it into standard language.”
korean-electronics.com | Blog Magazine of korean electronics, brands and Goods
Korean-Electronics.com 
Leave a comment