This Startup Lets You Search Video Like You Search The Web
These Columbia PhD students are revolutionizing the way video lives on the Internet
Rough Draft Ventures is excited to support Dan Morozoff and Joe Ellis of Vidrovr, a video processing platform better organizing video content. Vidrovr’s automatic systems index, tag, and understand video to let you search video like you search the web.
Currently, videos have to be manually tagged in order to be organized both online and offline. Not only is this process time consuming, it puts huge limits on video content’s reach and applications: Videos are currently tagged as a whole, which encourages generalized tags. This content is impossible for even modern search engines to find, and if it is discovered, there’s no understanding of what portions of the video are the most relevant or important to the query. Further, manual tagging is inherently subjective, making it difficult to build reliable recommendation and search systems around these descriptors.
Vidrovr leverages new advancements in deep learning and computer vision to revolutionize how video lives on the web. Vidrovr’s systems identify and leverage information like who’s talking, what text is showing on screen, and what sounds appear in the video to build a comprehensive understanding of the what the video is about. This multi-modal approach makes it the most sophisticated technology of its type.
This technology has widespread application across media, finance, and other verticals. Journalists, for example, can use Vidrovr’s Wordpress plugin to get video suggestions based on the text in their article. TV outlets can leverage Vidrovr to instantly and automatically produce specific video compilations like “all touchdowns Aaron Rodgers has made against the Bears.” There are opportunities for new products too — Facebook stockholders, for instance, may subscribe to a channel that bundles every clip that talks about the company.
The startup is currently working with one large media corporation and has five major trials.
Ellis and Morozoff met as PhD students at Columbia University in New York City, where Ellis was focused on video understanding and multimedia analysis, and Morozoff was focused on neuroscience. The two are currently taking a leave of absence to pursue Vidrovr full time.
Vidrovr went through The Combine incubator at the NYC Media Lab this spring, was recently awarded a $225K research grant from the National Science Foundation, and is a member of the Techstars NYC Winter Cohort.