Robert E. Wagner

Research

My current research interest is in real-time active vision.  Active vision systems require the use of efficient attentional operators for gaze control.  To gain insight into developing active vision algorithms, it is useful to study biological visual systems.  One such insight taken from the higher vertebrate visual systems is that they are space-variant.  In machine vision, space-variant image architectures provide a mechanism for data reduction and the possibility of real-time performance.

The main application of this research is face detection and tracking on a three degree of freedom stereo active vision system.  The attentional operators that are exploited fall along the visual dimensions of motion, color, depth and shape.  These operators are combined to locate and track the face.

In addition to this research, I am also interested in computer graphics and animation, as well as code optimization techniques.

Keywords: primate vision, real-time machine vision, robotics

Projects

	Color Tracking Demos View: colorTrack.mov (4 MB) robotColorTrack.mov (4.7 MB) rewColorTrack.mov (9.1 MB) These demos show that an inexpensive color filter can be used in a feedback loop to provide real-time control of the robot.
	Stereo Saccade Demo Demonstration of the result of a 3DOF calibration technique for pointing the cameras at an arbitrary point in the visual field. View: stereoSaccade.mov (7.3 MB)
	Stereo Motion Tracking Demo An alogorithm was developed to track a moving object with both cameras to obtain a rough estimate of range. The person is actually too far away from the camera to actually induce a visible vergence movement of the robot. View: stereoMotionTrack.mov (3.5 MB)
	Face Detection and Tracking Demo Note: The tracking portion of this demo is based purely on color segmentation. This movie illustrates how this can become problematic. View: faceTrack.mov (8.2 MB)
	Space-Variant Color Tracking Demo In this demo the forward monopole map is shown in the upper left corner and the inverse map is centered in the image. The largest color blob is depicted in red in the forward map and outlined with a blue rectangle in the inverse map. View: svColorTrack.mov (4.6 MB)

Publications

Wagner, R.E. (2004).  Stereo Active Vision and Peiripheral Optical Flow: Computer Vision Applications of the Wide-Field Human Visual Representation.  PhD thesis, Boston University, Boston, MA.

Gaudiano, P., Chang, C., Ecemis, I., Martens, S., Sahin, E. Streilein, W. and Wagner, R. (1999).  Neurobotics Lab Research: Learning, Vision and Sonar Recognition with Mobile Robots.  Neural Networks in Applications ’99, Fourth International Workshop, Otto-von-Guericke University of Magdeburg.

Sahin, E., Gaudiano, P. and Wagner, R.(1998).  A Comparison of Visual Looming and Sonar as Mobile Robot Range Sensor.  Proceedings of the Second International Conference on Cognitive and Neural Systems, Boston University.

Bacus, J.W., Bacus, J.V. and Wagner, R.E.   Method and apparatus for automated assay of biological specimens, US Patent 5,428,690, to Becton Dickinson and Company, Patent and Trademark Office, Washington, D.C., 1992.

Bacus, J.W., Bacus, J.V. and Wagner, R.E.   Method and apparatus for automated assay of biological specimens, International Patent 93102433.5, to Becton Dickinson and Company, 1992.

Cirruculum Vitae

My CV is available in the following formats: HTML, DOC, PDF, PS.