WFU Physics Colloquium

Title: "Tuned-Aperture Computed Tomography: a task-adaptable, three-dimensional imaging method."

Speaker:

Professor Richard Webber
Departments of Dentistry and Radiology,
Wake Forest University School of Medicine

Time:

4 PM, Thursday, Jan. 15, 1998

Place:

Room 101, Olin Physical Laboratory

Abstract:

Most practical noninvasive methods for generating three-dimensional (3-D) images such as computed tomography, holography, stereoscopy, tomosynthesis, magnetic resonance imaging, etc. require continuous stabilization of the projection geometry of the image-forming components during the period of data acquisition. This often constrains the range of applications involving objects that cannot be stabilized sufficiently. For example, living organisms from microbes to humans ultimately tend to move unpredictably during imaging exposure. Similarly, movement of inanimate objects as they progress along a conveyor belt or other non-calibrated production-based transportation device often precludes acquisition of 3-D information in a cost-effective manner.

A relatively simple method for acquiring 3-D data under these difficult conditions has been developed. It involves retrospective analysis of any number of two-dimensional projections produced from any number of unknown projection geometries. The method is perfectly general in that it can be applied to any projection-based imaging system, and it can make use of any kind of radiation including x-rays which cannot be easily focused using conventional technologies. The method parallels much of the flexibility afforded by existing optical systems. The latter often exploit manipulation of numerical aperture to facilitate control of depth of field and angular disparity associated with task-dependent display. However, the new method uses digital sampling methods to circumvent limitations imposed by lenses required to focus the radiation. This permits intermittent acquisition of data in short bursts with no need for restriction of object motion between bursts.

This new approach is called TACTã, an acronym for tuned-aperture computed tomography. A visual presentation will be used to demonstrate a conceptual basis for its historical evolution, to describe its algorithmic roots, and to offer some preliminary data to support its practical advantages. A real-time, computer-based demonstration will follow which will provide additional insight as to how this approach can be tailored to any number of task-specific applications.