(image: Pixabay)
SEOUL, April 17 (Korea Bizwire) — Korean researchers have made significant strides in video artificial intelligence (AI) technology, specifically in the area of sound-to-image conversion.
A research team led by Pohang University of Science and Technology (POSTECH) professor Oh Tae-hyun has developed a technology called Sound2scene, which can convert sound information into corresponding images.
The researchers analyzed hundreds of thousands of YouTube clips to develop the technology, and it has been named the world’s first recognizable sound-image conversion technology.
The team also developed “Sound to Text” technology, which can describe and read sound as text, and it has the potential to aid hearing-impaired people.
Additionally, the team has created an AI model that produces synthetic data using a small amount of actual data, allowing for better AI models.
This technology is being piloted at one of steel giant POSCO’s plants, and Prof. Oh believes it can improve the cognitive ability of robots and enhance the efficiency of the AI industry.
Kevin Lee (kevinlee@koreabizwire.com)
