Determining Optimal Textural Features for Cervical Cancer lesions using the Gaussian Function
S. Pradeep Kumar Kenny, S. Allwin · 2014
Cervical cancer is one of the most deadliest gynaecologic cancer occurring in women and it is also a very important research area in image processing. The main reason as to why this cancer is so deadly is that it cannot be detected early as it doesn't throw any symptoms until it evolves to the final stages of cancer. The reasons attributed to this is the cancer itself and also to the lack of pathologists available to screen the cancer especially in third world countries. One solution would be to automate the system. But this comes with a set of problems as we have a number of features to determine whether the given cell is cancerous or not. If we calculate with too many features then the accuracy goes up but speed comes down. And if you reduce the features it is the other way round. To strike a balance between the speed and accuracy of a system in this paper we have proposed a unique ranking system using Gaussian function to determine the right balance of features which can be used in an automated system.