Automated Segmentation of Cervical Cells Using MSER Algorithm and Gradient Embedded Cost Function-Based Level-Set Method

Abstract

Traditionally, cervical cells are screened by analyzing Pap smear slides. But this manual inspection requires expert pathologist making the entire process time consuming and prone to manual errors. Thus, it is needed urgently to develop an automated system for the screening process. Though extensive research work is going on for decades to develop the automated system, but the success is quite less owing to the fact in Pap smears, nuclei and cytoplasm are often found in clumps lacking any boundaries separating them. In this work, we have proposed a gradient embedded cost function for cytoplasm segmentation. We have used ISBI-15 dataset for the work and the result obtained is compared with the state-of-the-art techniques.