Columbia startup Vidrovr announced in December that it has secured a National Science Foundation Small Business Research grant for $225,000 to conduct research & development work on autonomous multimedia video understanding systems. The Vidrovr team was co-founded by PhD candidates Joseph Ellis ’17CU and Daniel Morozoff ’16CU.
Vidrovr developed an award-winning, patented news and social media video processing system called News Rover, which indexes, tags, and understands video content in real-time. News Rover analyzes multiple elements of video (such as pixels, audio, and text) and applies machine learning algorithms to identify people and objects on-screen, add tags, and link on-screen text to other text- and image-based information online.
News Rover technology is being leveraged in collaboration with enterprise customers and news media experts to design pilot programs that address clear market needs: automatic metadata generation for videos, video content management solutions that enable automatic placement, and recommendation of video clips for digital products. It also has the capacity to automatically link and source visual social media content that is relevant to a particular video or online article before it is published.
Ultimately, Ellis and Morozoff hope to develop a video search engine platform as well as a multimedia linking system that connects video to other digital content, thereby enabling creators to share and monetize video content online. “With this SBIR grant, Vidrovr will be able to explore unique ways to combine information from humans interacting with our system and the plethora of metadata extracted from multiple modalities associated with the video content into a unique intelligent learning framework, ” said Vidrovr CEO Joseph Ellis. “I am excited about the possibility of leveraging the expert knowledge within media companies and massive media data available to push the boundaries of the way computers can understand video.”
To learn more about the NSF SBIR/STTR program, visit: www.nsf.gov/SBIR.
To learn more about Vidrovr, visit their website.