Today video is everywhere, be it from surveillance cameras, TV or movies. However, most of the video data out there are unstructuralized. Structuralizing this data not only makes video contents more searchable but also carries additional benefits.
Today video is everywhere, be it from surveillance cameras, TV or movies. However, most of the video data out there are unstructuralized. Structuralizing this data not only makes video contents more searchable but also carries additional benefits.
That’s according to Viscovery, a start-up who entered the area of video recognition technology after realizing its potential. Today, the Taipei-based company is a leading provider of such technology in the Asia-Pacific region, with offices in Shanghai, Beijing and Singapore.
According to Amos Huang, CEO of Viscovery, the problem of video content has been the unstructured nature of it. “If you want to find a particular footage, for example, nowadays if you don’t have them previously tagged manually, it’s literally impossible to go through them. Even though there are certain video search engines, they are just a manually tagged result where the limit of it is based upon how many tags you put in,” he said. “There is no programmatic deep learning process for that. You cannot really search a video content, and that’s why we started to focus on video recognition technology.”
Viscovery’s patenting algorithm is called FITAMOS, which denotes the seven “things” that it analyzes: face, image, text, audio, motion, object and scene. It automatically processes video content into searchable, structured data. “So far, Viscovery has analyzed the TV series and movies that are popular these three years, which make them capable of being immediately recognized by every scene, the period of appearance, and the proportion of the screen where specific celebrities, commodities, or brands appear. By doing so, users can really search the video content instead of just searching the key words of the title or introduction,” Huang said.
According to him, this opens up new opportunities, especially in the area of contextual advertising.
“Google AdSense has already proved that contextual ads are effective; when the text provides related information at the right place and the right time, the users’ click through rate (CTR) will increase. It also works on video content; after their online test, they found that video contextual ads have in average 32 percent more CTR than traditional video ads, and it even had an 86 percent growth on tablets,” Huang said. “Viscovery’s video data analysis can effectively find the opportunity or the best scene to put advertisements for any kind of commodity or brand. The ads suitable for the scene are then triggered to provide seamless ad experience to audiences.”
Viscovery’s technology can have security applications, too. “Our technique can analyze video and detect whether there is an intruder or a car parked illegally, but we are more focusing on contextual ads and video search engine right now,” Huang said.