Segmentation Method for Nuclei Cell Image Using Hybrid Techniques
In the medical applications, the nuclei images are the most important for classifying the tumors into normal or abnormal (benign and malignant). The important stride in image histometric and cytometry is an automatic segmentation and clustering of cell nuclei. Despite substantial progress, there is a requirement to enhance speed, precision, automation level, and adaptation to modern implementations. A cancer diagnosis for early, cell nuclei automated segmentation is decisive such as characteristics of the nucleus of the cell are fundamentally related to an evaluation of maliciousness. Very little research work was done implemented in cell nuclei segmentation of automated on pleural effusion images cytology, that is seedy treated through previous techniques. Furthermore, cytology pleural effusion image is still defying because of an assortment of cells, the poor contrast of images, and overlapping cells. The recent cancer of cervical diagnosis, Pap smear examination rolls an essential role in that cells of the human obtained of the patient of the cervix are analyzed of precancerous modification. The cells expert is performing the manual analysis of these cells and this work takes time-consuming and labor intensive. In this paper, we present a novel approach of obtaining a Nuclei cell segmentation method using a hybrid technique and evaluate this technique using the Mean Square Error (MSE) between the centroid from image segmentation and its original label. Two types of image data sets, benign and malignant images, are implemented in this work. The result obtained in this work shows a confident approach of segmentation for nuclei cell image. The average rational error of the MSE obtained in this work is about 11.361 %.