Journal of Pure and Applied MicrobiologyVol. 9 No. 4

Segmentation and Abnormality Detection of Cervical Cancer Cells Using Fast Extreme Learning Machine with Particle Swarm Optimization

V. Sivaraj1, S. Sukumaran1 and P.Sukumar2

1Department of Computer Science, Erode Arts and Science College, Erode - 639 009, India. 2Faculty of Electronics and Communication Engineering, Nandha Engineering College, Erode, India.

Received on 14 September 2015 and accepted on 12 November 2015

 

ABSTRACT

Cervical cancer occurs only when the abnormal cells on the cervix will mature and unable to manage clearly in the reformation area. Mostly used technique for detecting the abnormal cervical cells is the routine and there will be no dissimilarity among the normal and abnormal nuclei. The color which is brown is abnormal nuclei and blue is the normal nuclei. Based on the Iterative Decision Based Algorithm, the cells are examined and the denoising of images is performed. Segmentation of the image is the procedure of grouping the digital image into compound sections. The preceding technique namely Support Vector Machine (SVM) will able classify only few nuclei regions but it will take high execution time. So, this research proposed a method called Fast Particle Swarm Optimization with Extreme Learning Machine (Fast PSO-ELM) for classifying all regions of nuclei into touching and non-touching region. This method is more efficient when compared with SVM method.

Keywords : Cervical Cancer, Image Denoising, Extreme Learning Machine, White Blood Cells, Particle Swarm Optimization, Fast Extreme Learning Machine.