An expert system for control and cell classification

The objective of this work was to develop and test rule-based expert-system techniques to control an interactive cytophotometry system. The application for this system was the detection and measurement of individual cells stained specifically for the human immunodeficiency virus (HIV) in patients with AIDS. A microscopic photometry system was used to acquire digitized images of labeled cells. The expert system consisted of a set of program modules that took the digitized video image through several data transduction phases to the final state of identification and measurement. Human experts interactively developed rules to identify cells, determine the cell type, and determine the staining density and patterns of individual cells.<>