• 
• 
• 
• 
• 
• 

Event Detection

The Murray Hill Laboratory (MHL) of Mitsubishi Electric Research Laboratories has developed a suite of Video Indexing, Summarization and Segmentation techniques. We have obtained promising results by applying them to event detection in traffic video and soccer video. We have obtained promising initial results and now plan to develop more accurate and sophisticated techniques for event detection in traffic and indoors surveillance video. The evident application is in the area of surveillance such as detecting accidents through analysis of traffic surveillance video illustrated at left. We have promising preliminary results on detection of onsets of traffic jams using our MPEG-7 motion activity descriptor. In another project, we are jointly investigating extraction of semantic features from low-level features of soccer games with Columbia University.

Background & Objective:  The proposed MPEG-7 standard will standardize a multimedia content description interface that will enable efficient searching and browsing of worldwide multimedia content. In this project we will work on both analysis of stored video using searching and browsing techniques for event detection as well as on direct event detection using MPEG-7 based techniques for content understanding in combination with our video segmentation technology.

Technical Discussion:  Our system is currently based on dynamic feature extraction in the compressed domain. It primarily relies on the MPEG-7 motion activity descriptor, and also makes use of simple color histograms. We have a unique motion activity based approach to video summarization. We plan to refine our techniques as well as customize them to traffic video analysis.

Technology Areas:
Audio Video Processing
Computer Vision

Modification Date:  June 13, 2008