Available TechnologiesInformation for InventorsInformation for IndustryUSM PoliciesRelated SitesPresentationsFAQsStatistics Start-ups About OTCSite IndexNews & Events
Home > News & Events > Featured Technology > Archived Featured Technologies > Shot Boundary Detection and Keyframe Extraction Using Pixel-To-Neighbor Image Differences < You Are Here

Technology:

Shot Boundary Detection and Keyframe Extraction Using Pixel-To-Neighbor Image Differences

Reference: 
IS-2005-104
Inventor: 

Dr. David S. Doermann and Ming Luo

Video analysis plays an important role in multimedia and surveillance industries. An essential component of video analysis is shot boundary detection. This fundamental step is included with low-level feature extraction, high-level feature extraction, and search (interactive, manual and automatic) in video retrieval and evaluation technologies.

University of Maryland researchers have developed novel methods for shot boundary detection of cuts, dissolves and fades. These new software algorithms provide improved accuracy for a wider variety of shot changes. Better video summarization, indexing and retrieval tools can be created from this technology. This technology is licensed to a start-up company.

 

This page is maintained by the Office of Technology Commercialization
Copyright © 2006. University of Maryland. All rights reserved.
Contact OTC
This page was last modified March 9, 2006 1:54 PM .