Performance analysis of histogram-threshold method for cancer detection

Abstract

In this paper, histogram-threshold method developed for cancer detection using miniaturized confocal microscopy system and its related performance analysis are presented. While doing the performance analysis, the receiver operation characteristics are applied in a novel fashion. Additionally, noise performances of different method are investigated. The highest and lowest success rates with 91.67% and 27.08% are acquired using Entropy Method and Mean Method, respectively.