Feature set for extraction of cervical cyto images

Cervical cyto images have found keen interest among research enthusiasts in the recent past, not only because of its killer reputation but also due to the diverse challenging unsolved problem it possess. Cervical cancer like all other cancer develops through various stages before it actually causes potential harm. These stages can be detected through visual information such as the shape, texture etc., by a pathologist. However there aren´t so many pathologists around to do the job leaving the alternative choice to an automated system to detect these features. Here we have set out to detect these features. The feature set chosen by us comprises of three different techniques which detect the cancer at various stages. The final outcome detects the cancer more efficiently and shows an accuracy of 87.67%. This work will surely help the pathologist detect the cancer more efficiently.