Krishna Gopal Dhal

Name: Krishna Gopal Dhal

Affiliation: Assistant Professor, Midnapur College

Title: Partitional Clustering based Fast and Robust Digital Pathology Image Segmentation

Area: Biomedical Signal and Image Processing

Abstract: Accurate segmentation of different regions of digital pathology images is of the utmost importance for a Computer Aided Diagnosis (CAD) system. Pathology image segmentation techniques must be fast, accurate, and noise-robust. Partitional clustering techniques like K-Means (KM) and Fuzzy C-Means (FCM) can be enhanced to fulfil the mentioned needs. Image histogram-based or superpixel image-based clustering techniques can be very useful to perform pathology image segmentation. Finding and representing the histogram of colour images is a hard task, and classical histogram-based clustering techniques are not noise-robust. Hence, superpixel image-based clustering techniques currently attract the interest of researchers. A superpixel image helps to reduce the execution time of the clustering technique and also has noise-resistance. The local optima trapping problem of the KM and FCM can be overcome to some extent by using nature-inspired optimizers (NIO) like particle swarm optimizer (PSO). Recently, researchers have also been working in this domain, i.e., NIO-based clustering techniques for digital pathology image segmentation.