Fractional-Order Darwinian Swarm Intelligence Inspired Multilevel Thresholding for Mammogram Segmentation
A. Santhos Kumar, Ata. Kishore Kumar, Varun Bajaj, Girish Kumar Singh · 2018
Multilevel thresholding is more accurate than the classical thresholding method for segmenting a digital mammogram, since it uses more number of intensities to represent the objects. It is intuitively appreciable for further detection of breast cancer using such kind of clinical database. The main goal of any multilevel thresholding based segmentation is to optimize its objective function to obtain different threshold levels; but multilevel thresholding is computationally expensive, and sometimes these optimized values are not accurate. Therefore, in this paper, an amalgamation of n-level thresholding and fractional-order Darwinian particle swarm optimization (FODPSO) is studied in detail, and was found to be the best among various PSO variants based thresholding for efficient mammogram image segmentation. The efficiency of the proposed technique is compared with other segmentation technique, based on thresholding such as particle swarm optimization (PSO) and Darwinian particle swarm optimization (DPSO). Individual performances of hereby employed algorithms are compared and analysed using the performance measures such as PSNR, SNR, SSIM, and MSE.