Optimal image segmentation of cancer cell images using heuristic algorithms

Abstract

In this world, protection of health from diseases is quite challenging. Cancer is one of the most harmful diseases which pose a major threat to human. There are two types of cancer tumours developed in human tissues namely benign and malignant. A benign tumour is a mass of cells that lacks the capacity to invade neighbouring tissue or metastasize. A malignant tumour is developed from benign tumour by the process called as tumour progression. This tumour invades neighbouring tissues rapidly and causes organs to get malfunction. In this paper, two benign and malignant images (512 × 512) are taken and evaluated using heuristic algorithms, such as PSO, DPSO, and FODPSO algorithms existing in the literature. The proposed segmentation procedure is executed using the conventional Otsu’s between-class variance function. The performances of considered algorithms are analyzed using the popular image parameters, such as objective value, Root Mean Square Error (RMSE), and Peak Signal to Noise Ratio (PSNR), and number of iterations. Results of this study demonstrate that FODPSO offers better result compared to PSO, and DPSO algorithm.