Lee Gyu-lee is a business writer at The Korea Times, focusing primarily on IT & telecommunications, the Ministry of Trade, Industry and Energy and KOTRA. Prior to this, she has covered a wide range of cultural news, from film, television and K-pop to lifestyle and fashion.
Naver Cloud unveils advanced AI video analysis
Naver Cloud demonstrates its new AI video analysis assitant, Media AI Understanding, at Korea International Broadcasting, Media, Audio, and Lighting Show in southern Seoul, Wednesday. Courtesy of Naver Cloud
By Lee Gyu-lee
Naver Cloud has unveiled its new AI video analysis assistant, Media AI Understanding (MAIU), demonstrating it at this year’s Korea International Broadcasting, Media, Audio, and Lighting Show, taking place from Tuesday to Friday in southern Seoul.
The service automatically analyzes various types of information from videos, including people, actions, backgrounds and audio with AI to generate metadata. It allows users to search for specific scenes with simple text searches, such as searching for scenes with the keywords “wine glass" or "male lead character running.”
"The rapidly growing video industry is one of the most anticipated areas for innovations through the adoption of cloud and AI technology," the company said.
"We will strive to meet various consumer needs and lead innovation in the video industry with end-to-end media services that encompass the entire production process from content production, editing and distributing to management, along with our AI video analysis services."
The new AI assistant is expected to help production studios, broadcasters and other content creators to majorly reduce their turnaround time for production.
By using MAIU, users can dramatically save time in extracting metadata, which normally requires the most time in production, and conduct preliminary editing more quickly.
Typically, one episode of a variety show uses approximately 3,600 minutes of recorded footage, taking about three to four days, or about 32 hours, to go through the footage for the preliminary editing process. But, with MAIU, the task can be completed within two hours, about 15 times quicker than it previously took.
The currently existing AI analysis services are generally limited to detecting or recognizing objects by individual frames, making it difficult to grasp the overall flow or context of video content. However, Naver Cloud explained that MAIU captures segments of context through analyzing by scenes rather than single frames, allowing it to provide a more intuitive structure for users to utilize in professional editing and reprocessing.
The service is currently going through closed beta testing with local broadcasting companies and partners, and is scheduled to officially launch next month.
Its price will be set at up to one-third of other global competing products to make it more approachable and price-efficient. The users can also distribute their content through mobile live streaming and video-on-demand within the same platform, enabling efficient processes from production to distribution.
NAVER Cloud also introduced various media services at the exhibition, showcasing features such as real-time translation and subtitles, beauty and avatar filters and automatic short clip generation. The company is also set to attend Media Technology Strategy, scheduled for Thursday.