An interactive image analysis system for quantitative cytology & to classify cervical cells.

Abstract

Introduction of computers and image analysis systems are gaining faster momentum in order to quantitate the assessment of cells for diagnosis and prognosis, and this system aims to relieve the operator from the tedium of microscopic observation and reduce operator bias and human error. This paper discusses the design and configuration of an interactive image analysis system built in the laboratory for the purpose of cell analysis and classification. The software developed to compute various textural and morphological parameters of cells on smear are briefly described. The results of the experiments carried out to classify normal and abnormal cells on cervical smear show 94 per cent success rate.