Google Research Award to improve video searches

August 02, 2011
Print

Greg Mori, associate professor with SFU Computing Science, and his PhD student Weilong Yang will work with Google researchers to refine video searches under a new one-year Google Research Award.

 

Mori will develop computer vision algorithms that will automatically analyze Internet video data, such as YouTube videos, and generate a set of tags (otherwise known as search terms). Tags are used today for searching video content, but they have the potential to be more precise and applied to sophisticated recall, browsing and analysis of video content. For example, Mori aims to more precisely apply tags to specific regions of a video, allowing thumbnail views or visual summaries to be produced that pinpoint content within a section. This will enable someone searching for information to browse and find the exact region of a video that is relevant.

 

This technology could be applied to enhance YouTube’s automatic tag suggestion in place today. A subsidiary of Google, video sharing website YouTube is the world’s largest with an estimated 490 million users worldwide (unique visitors per month) and an estimated 92 billion page views each month.

 

The research award from Google recognizes the growing need to retrieve and organize video information more effectively.

 

For more information, contact fascomms@sfu.ca.

No comments yet