Multi Class Cervical Cancer Classification by using ERSTCM, EMSD & CFE methods based Texture Features and Fuzzy Logic based Hybrid Kernel Support Vector Machine Classifier
S. Athinarayanan, Dr.M.V. Srinath, R. Kavitha · IOSR Journal of Computer Engineering · 2017
Cervical cancer is the highest rate of incidence after breast cancer, gastric cancer, colorectal cancer, thyroid cancer among all malignant that occurs to females ; also it is the most prevalent cancer among female genital cancers.Manual cervical cancer diagnosis methods are costly and sometimes result inaccurate diagnosis caused by human error but machine assisted classification system can reduce financial costs and increase screening accuracy.In this research article, we have developed multi class cervical classification system by using Pap Smear Images according to the WHO descriptive Classification of Cervical Histology.Then, this system classifies the cell of the Pap Smear image into anyone of five types of the classes of normal cell, mild dysplasia, moderate dysplasia, severe dysplasia and carcinoma in situ (CIS) by using individual and Combining individual feature extraction method with the classification technique.In this paper three Feature Extraction methods were used: From that three, two were individual feature extraction method namely Enriched Rough Set Texton Co-Occurrence Matrix (ERSTCM) and Enriched Micro Structure Descriptor (EMSD) and the remained one was combining individual feature extraction method namely concatenated feature extraction method (CFE).The CFE method represents all the individual feature extraction methods of ERSTCM & EMSD features are combining together to one feature to assess their joint performance.Then these three feature extraction methods are tested over Fuzzy Logic based Hybrid Kernel Support Vector Machine (FL-HKSVM) Classifier.This Examination was conducted over a set of single cervical cell based pap smear images.The dataset contains five classes of images, with a total of 952 images.The distribution of number of images per class is not uniform.Then the performance was evaluated in both the individual and combining individual feature extraction method with the classification techniques by using the statistical parameters of sensitivity, specificity & accuracy.Hence the resultant values of the statistical parameters described in individual feature extraction method with the classification technique, proposed EMSD+FLHKSVM Classifier had given the better results than the other ERSTCM+FLHKSVM Classifier and combining individual feature extraction method with the classification technique described, proposed CFE+FLHKSVM Classifier had given the better results than other EMSD+FLHKSVM & ERSTCM+FLHKSVM classifiers.