DoD TechMatch: SBIR Opportunity View

Site Search:

Print This Page

	STTR Opportunity View

Back to Listing

ST081-010




Go To DOD SBIR Site
Basic Information:

Title:	Combat Video Analysis Engine
Program:	STTR
Technology Area:	Information Systems, Sensors
Open Date:	2/19/2008
Close Date:	3/19/2008

Description:

Although massive quantities of aerial and ground video are available from military operations, automatic analysis of this data is generally limited to detection and tracking of objects and simple event recognition. The challenges preventing the automated detection of activities include huge amounts of background clutter, handling fragmented tracks, large variations in activity configurations, and effectively representing a wide variety of activity models in a computable form. Furthermore, the algorithms must operate on low resolution video (greater than 10 cm per pixel) from which identification of individuals is not possible. DARPA is seeking methods to detect threats against US forces that can overcome these challenges while processing full motion combat video in real time. Activities of interest should be detected without assuming that objects are identified or accurately tracked through the duration of the activities. Massive amounts of data should be rapidly filtered to focus resources on likely activity candidates early in the processing chain. Activity models should handle partial temporal ordering, variable numbers of objects, and significant variations in activity duration. Metadata from sensors, platform, and from mission plans should be exploited. Contextual data such as terrain models, 3D urban models, road networks, cultural metadata, observed traffic patterns and other information should be used to reduce false alarms and resolve ambiguity.

Objective:

Develop new tools for activity analysis of combat aerial video collected for tactical military operations. These tools should leverage new methods in computer vision, machine learning and probabilistic models to detect and recognize complex threats and suspicious activities without identification of specific individuals.

Phase I:

Develop activity model representation, and overall approach to scaling up to massive video streams of military significance. Perform initial feasibility studies to estimate the scalability and accuracy of the approach.

Phase II:

Develop and implement the complete system architecture. Compare with existing methods to demonstrate improvement over the state-of-the-art on large, realistic data sets. Develop an operational prototype, or integrate the developed tools into an existing government system.

Phase III:

Candidate applications of this technology span both military and commercial requirements. In general, the availability of image and video data is growing substantially. Commercially, the proliferation of security cameras for a variety of purposes, such as traffic monitoring, yields fertile territory for a technology which can support detection and analysis of activities of interest. This capability will support both post-event forensic analysis and monitoring of live feeds to give alerts on immediate activities of interest, such as an automobile accident. This capability also has applicability to other applications, such as web-based video searches and video-library searches.

References:

1. Gong and T. Xiang. Recognition of group activities using dynamic probabilistic networks. In Proceedings of the International Conference on Computer Vision, 2003, Vol. 2, p. 742. IEEE. 2. S. Hongeng, R. Nevatia, and F. Bremond. Video-based event recognition: activity representation and probabilistic recognition methods. Computer Vision and Image Understanding, 96(2):129162, November 2004. 3. M.T. Chan, A. Hoogs, R. Bhotika, A.G.A. Perera, J. Schmiederer and G. Doretto. Joint Recognition of Complex Events and Track Matching, Proceedings of the Conference on Computer Vision and Pattern Recognition, Vol. 2, 2006, pp. 1615-1622. IEEE. 4. Y. Shi, Y. Huang, D. Minnen, A. Bobick, and I. Essa. Propagation networks for recognition of partially ordered sequential action. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Vol. 2, 2004, pp. 862-869. IEEE. 5. N. Dimitrova, H. Zhang, B. Shahraray, I. Sezan, T. Huang, and A. Zakhor, Applications of video content analysis and retrieval, IEEE Multimedia, vol. 9, no. 3, pp. 42-55, JulySept. 2002. 6. Digital Video Retrieval at NIST, http://www-nlpir.nist.gov/projects/t01v/

STTR Keywords

content-based video indexing and retrieval, content-based image indexing and retrieval, content-based multimedia indexing and retrieval, activity-based video indexing and retrieval, intelligent retrieval of surveillance video, similarity-based video retrieval, activity recognition, search by example

TechMatch Keyword(s):

Computer Software

Applications

Data Fusion

Data Visualization and Graphics

Image Processing and Pattern Recognition

Information Processing

Knowledge Management

Signal Processing

Geographic Information Systems (GIS)

Decision Support

Databases

Electronics and Electrotechnology

Displays

Mathematical Sciences

Algorithms

Operations Research

Sensors

Training and Education

Simulation

Wargaming

Virtual Reality

Imagery

Modeling

Engineering

Systems

Urban Operations

DISCLAIMER:

The Solicitations listed on this site are copies from the various SBIR agency solicitations and are not necessarily the latest and most up-to-date. For this reason, you should use the agency link listed below which will take you directly to the appropriate agency server where you can read the official version of this solicitation and download the appropriate forms and rules. The Official Website can be found at: http://www.acq.osd.mil/sadbu/sbir/solicitations/index.htm