Abstract:

Medical images today help bring to divulge hitherto concealed knowledge about diseases that were, at one time, rarely subject to intense and prolonged scrutiny and analysis. In recent times, however, imaging has gone a long way in helping establish plausibly both the causes and behavioural patterns of a given disease. The objective of this paper is to constitute a series of techniques to detect accurately, in lymphocytes from blood smear images, the occurrence of Rheumatoid Arthritis (RA). This paper has used computer-aided diagnosis for accuracy and consistency, and the threshold segmentation method with slider control as a proposed segmentation method for lymphocyte extraction from blood smear images, precisely because it performs much better than existing segmentation methods. The paper discusses critical medical parameters - such as area, perimeter, circularity, roundness and solidity - from segmented lymphocytes, and also describes the ADTree method governing classification and decision rules. The final part of the paper deals with a case study on datasets of inflamed and non-inflamed types (for three different medical cases) using correlation and Analysis Of Variance (ANOVA) techniques to discover the homogeneity and relationships that exist between the critical parameters listed above for identifying the status of RA.

Keywords: