Riad Hammoud
Delphi Corporation,
Delphi Electronics and Safety
World Headquarters, IN, USA
Tel/Fax: 1-765451-9802/1340
riad.hammoud@delphi.com

Workshop Co-Chair:

Robert W. McMillan
US Army Space and Missile Defense Command
Huntsville, AL, USA
bob.mcmillan@us.army.mil

Benchmark and Website Chair:

James W. Davis
Ohio State University, USA
jwdavis@cse.ohio-state.edu

Demo and Poster Chair:

Longin J. Latecki
Temple University, USA
latecki@temple.edu

Program Committee:

Mohamed Abdel-Mottaleb
U of Miami, USA

Besma Abidi
U of Tennessee, Knoxville, USA

Gregory Baratoff
SiemensVDO Automotive,
Germany

Mike Bazakos
Honeywell Labs, USA

George Bebis
U of Nevada, Reno, USA

Bir Bhanu
U of California, Riverside, USA

Patrick Bouthemy
INRIA/IRISA, France

Alberto Broggi
Universita di Parma, Italy

James W. Davis
Ohio State U, USA

Larry Davis
U of Maryland, USA

Riad Hammoud
Delphi E&S, USA

Katsushi Ikeuchi
IIS, U of Tokyo, JAPAN

Longin Jan Latecki
Temple U, USA

Tanveer Syeda-Mahmoud
IBM Almaden Research Center, CA, USA

Robert McMillan
U.S. Army Space & MDC, USA

Swarup Medasani
HRL Laboratories, CA, USA

Gerard Medioni
U of Southern California, USA

Nasser Nasrabadi
Army Research Lab, USA

Ioannis Pavlidis
U of Houston, USA

Firooz Sadjadi
Lockheed Martin Corp, USA

Andrea Selinger
Equinox Corp, USA

Mubarak Shah
U of Central Florida, USA

Tan Tieniu
CASIA, CHINA

Mohan Trivedi
U of California, San Diego, USA

Djemel Ziou
U of Sherbrooke, Canada

Industrial Liaison

Andrew P. Harbach
Delphi Corporation, USA

Sponsors:

Questions?

riad.hammoud@delphi.com

The scope of Object Tracking and Classification Beyond the Visible Spectrum workshop series encompasses many disciplines, including infrared, far infrared, millimeter wave, microwave, radar, and synthetic aperture radar sensors as well as the very dynamic topics of signal processing, computer vision, and pattern recognition. It is a fertile area for growth in both analysis and experimentation and includes both civilian and military applications. The availability of ever improving computer resources and continuing improvement in sensor performance has given great impetus to this field of research. This technology "push" has been balanced by a technology "pull" resulting from increasing demand from potential users of this technology including both military and civilian entities as well as needs arising from the growing field of Homeland Security.

The original tentative of the vision community has been focused mostly on the development of vision algorithms for object detection, tracking, and classification associated with visible range sensors in day and office-like environments. In the last decade, infrared, thermal and other non-visible imaging sensors were used only in special areas like medicine and military. That lower interest level in infrared imagery was due in part to the high cost of non-visible range sensors, low image resolution, high image noise, lack of widely available data sets, and lack of consideration of the potential advantages of non-visible lights. These historical objections are becoming less relevant as infrared imaging technology advances and their cost is dropping dramatically. Image sensing devices with high dynamic range and high IR sensitivity have started to appear in a growing number of applications ranging from military and automotive domains to home and office security applications.

In order to develop robust and accurate vision-based systems that operate in and beyond the visible spectrum, not only existing methods and algorithms originally developed for the visible range should be improved and adapted, but also entirely new algorithms that consider the potential advantages of non-visible ranges are certainly required.  The fusion of visible and non-visible ranges, like radar and IR images, or thermal and visible spectrum images, is another dimension to explore for a higher performance of vision-and-signal based systems. The non-visible light is widely employed in night vision-based systems, and many detection and recognition systems available today in the market are relying on physiological phenomena produced by IR and thermal wavelengths. Using artificial controlled lights is a practical solution to eliminate challenging ambient light effects.

This series of OTCBVS workshop creates connections between different communities in the machine vision world ranging from public research institutes to private, military, and medical laboratories. It brings together pioneering academic, industrial and military researchers and engineers in the field of computer vision, image analysis, pattern recognition, signal processing, sensors, and human-computer interaction.

Topics of Interest:
Special attention will be given to vision algorithms where non-visible sensors are employed. However, we also encourage the submission of high quality papers that deal with object tracking and classification in the visible spectrum. Additionally, emphasis will be placed on new and traditional applications of visible and non-visible imagery.

In this second annual meeting of OTCBVS workshop series, a benchmark/test dataset of images and videos recorded in and beyond the visible spectrum is available: 

The dataset is to be used to compare, evaluate, and adapt state-of-the-art computer vision algorithms. It will fill out the lack of experimental non-visible data in the vision community, and will allow a large spectrum of CVPR participants to explore the benefits of the non-visible spectrum in real-world applications, and contribute to OTCBVS workshop series. We also invite people to participate in the extension of this dataset.

Comparative evaluation studies across the non-visible spectrum for a given computer vision or pattern recognition task are encouraged. Applications using non-visible sensors from various domains are welcome. Sensors of interest include visible, infrared, millimeter wave, radar, and hyper-spectral. The topics of OTCBVS'05 include:

Other topics dealing with non-visible sensors are welcome. The object category includes, but is not limited to eye, face, vehicle occupant, pedestrian, vehicle, landmine, ship, airplane, and missile.

Demo Session:
We also encourage submissions of system demos. The goal of our demonstration session is to demonstrate working vision systems in an environment that allows conference participants to view them first hand in a manner not possible in a technical talk or poster session alone. We invite proposals for demonstrations from anyone in the computer vision community, including researchers, academic institutions, and industry. Demonstrations submitted before April 10, 2005 will be allotted 2 PDF pages in the CDROM proceedings. Later demonstration submissions are possible, but will not appear in the proceedings.

Sponsorship & Best Paper Award:
The IEEE Computer Society, Delphi Corp, and US Army Space and Missile Defense Command will sponsor our workshop. Delphi will award the best paper presented at this meeting, and cover the expenses of the keynote speaker. The workshop will have its own printed proceedings, which will be distributed to all workshop attendees.

Past Meetings:
The first annual workshop took place in Washington DC on July 2, 2004 in conjunction with IEEE CVPR 2004. More details about the program, attendees, and papers are found at this web page:

Some papers of the workshop were pre-selected for publication in a special issue of International Journal of Computer Vision (IJCV). An open call for papers has been issued for this special issue of IJCV on OTCBVS. More details can be found at:

Paper Submission:
Papers are required by midnight March 27, 2005 EST. The submission is electronic, and must be in PDF format. Papers should not exceed 8 double-column pages. The paper format must follow the standard IEEE 2-column format of single-spaced text in 10 point Times Roman, with 12 point interline space. All paper submissions must be anonymous. See CVPR 2005 Author Instructions page for detailed guidelines.

If you experience a problem during the electronic submission, please contact the Workshop Chair.  Accepted papers will be published in the workshop proceedings and they will appear in the CDROM of the main CVPR conference as well.