CERVICAL CANCER DETECTION AND CLASSIFICATION BY USING EFFECTUAL INTEGRATION OF DIRECTIONAL GABOR TEXTURE FEATURE EXTRACTION AND HYBRID KERNEL BASED SUPPORT VECTOR CLASSIFICATION

Abstract
Planning of invigorating representation is a troublesome and testing process because of the unpredictability of the images and absence of models of the life systems that thoroughly catches the reasonable expressions in each structure. Cervical malignant growth is one of the noteworthy reasons for death among different kinds of the diseases in women around the world. Genuine and auspicious determination can keep the life to some dimension. Therefore, we have proposed a computerized dependable framework for the analysis of the cervical malignancy utilizing surface highlights and machine learning calculation in Pap smear images, it is extremely advantageous to anticipate disease, likewise expands the dependability of the determination. Proposed framework is a multi-organize framework for cell nucleus extraction and disease finding. To begin with, clamor expulsion is performed in the preprocessing venture on the Pap smear images. Exterior highlights are separated from these demand free Pap smear images. Next period of the proposed framework is classification that depends on these separated highlights, SVM classification is utilized. Over 94% exactnessis accomplished by the classification stage, demonstrated that the proposed calculation precision is great at recognizing the disease in the Pap smear images.

Authors
S Athinarayanan1, K Navaz2, R Kavitha3, S Sameena4
Annamacharya Institute of Technology and Sciences, India1,2, The MDT Hindu College, India3, Universal College of Engineering and Technology, India4

Keywords
Cervical Cancer, Feature Extraction, DGTF, Classification, Hybrid Kernel SVM
Published By :
ICTACT
Published In :
ICTACT Journal on Image and Video Processing
( Volume: 9 , Issue: 3 )
Date of Publication :
February 2019

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.