Video Analytics

Summary

The Multimodal Information Group's (MIG) video analytics program includes several activities contributing to the development of technologies that extract information from video streams through systematic, targeted annual evaluations and metrology advances. Two currently active evaluations are: Activities in Extended Video (ActEV), and the Nimble Challenge Evaluation for Media Forensics.

There has been a huge growth in video analytics research over the past decade. The MIG group saw this as an opportunity to apply our group's three decades of knowledge in language content recognition and extraction metrology to video analytics research. This includes defining performance metrics, building real-world data sets, and identifying key component technologies that must be developed to build a video analytics pipeline from the pixel level to the content understanding level.

Description

Active evaluations/projects include:

• Activities in Extended Videos: The ActEV evaluation seeks to evaluate robust automatic activity detection algorithms for a multi-camera streaming video environment. ActEV will address activity detection for both forensic applications and for real-time alerting.
• Media Forensics Challenge Evaluation: The objective of the Media Forensics Challenge Evaluation is to promote the development of technologies that automatically assess the integrity of an image or video determining if the media is authentic.

Past projects include:

• Multimedia Event Detection: The goal of MED is to assemble core detection technologies into a system that can search multimedia recordings for user-defined events based on pre-computed metadata. The metadata stores developed by the systems are expected to be sufficiently general to permit re-use for subsequent user defined Ad-Hoc events.
• Surveillance Event Detection:  The objective of the Surveillance Event Detection evaluation is to promote the development of technologies that detect activities that occur in the surveillance video domain.
• Video Surveillance Technologies for Retail Security: The focus of the VISITORS project was to advance predictive analysis technologies and methodologies that are able to detect persons engaged in suspicious activities in surveillance video. VISITORS is being applied in the retail domain.
• AVSS Multiple Camera Single Person Tracking (MCSPT) Challenge: The goal of the MCSPT evaluation was to facilitate research via a common evaluation task that focuses on one aspect of person tracking technologies: the ability to track a specified person within a video sensor field using a small set of in situ exemplar video images to specify the person.
• Video Analysis and Content Extraction (VACE): The US Advanced Research and Development Activity VACE program was established to develop novel algorithms for automatic video content extraction, multi-modal fusion, and event understanding. During the program, progress was made in the automated detection and tracking of moving objects, including faces, hands, people, vehicles, and text.
• Classification of Events, Activities and Relationships (CLEAR): CLEAR was a multi-national evaluation series that brought together the researchers from the US ARDA VACE Program and the European Union Computers in the Human Interaction Loop Program to focus research on detecting and tracking people, faces, vehicles, etc. and acoustic event detection.