Colloquium Details
Partial differential equations in interpretation of a class of ophthalmological images
| Author: | Andrew P. Paplinski Computer Science & Software Eng, Monash University, Melbourne, Australia |
|---|---|
| Date: | February 03, 2000 |
| Time: | 16:00 |
| Location: | 220 Deschutes |
Abstract
The results presented in this seminar originate from ongoing work on interpretation of a class of ophthalmological images, namely, the Posterior Capsular Opacification (PCO) images acquired in St. Thomas' Hospital, London after cataract operations. The work is conducted in co-operation with the Image Processing Group, Department of Physics, King's College London.
We start with the background information regarding acquisition of the images, their general characterization, and a formulation of a problem which we attempt to solve.
Then we briefly describe the current method of segmentation of the PCO images based on a directional variance operator and co-occurrence arrays. The method is now routinely used and works well in majority of cases.
In order to deal with difficult-to-segment images, we have investigated non-linear filtering method based on application of partial differential equations. We present results of application of two image processing/enhancement algorithms, namely, an anisotropic diffusion method and a curvature-driven min/max flow.