Detection of Cervical Cancer using GLCM and Support Vector Machines
Shalini Nehra, Jagdish Lal Raheja, Kaustubh Butte, Ameya Zope · 2018
Early detection of cancer can lead to a higher likelihood of survival, lower costs of care which would further result in lowering death rates and disability due to cancer. Introduced in March 1924, colposcopy was used to detect the cause of abnormal looking cervix and hence provide appropriate treatment. Being prone to human errors, colposcopy can lead to misleading results for detection of cervical cancer. This paper aims at presenting a convenient and automated method of detection of cervical cancer and classifying images as cancerous or non-cancerous. After applying scaling on the images obtained via colposcopy, the Gray Level Co-occurrence Matrix (GLCM) was constructed. The Haralick features extracted from the above-constructed matrix were hence used to classify images into cancerous and non-cancerous using Support Vector Machines (SVM).