HYBRID CLASSIFIER TO CLASSIFY THE FINGER NAIL ABNORMALITIES

Abstract

Nail diagnosis is a method to predict the possibilities of organ failures and various systemic diseases. Nail abnormalities are considered as the signs of certain diseases in traditional medicines such as Siddha Medicine, Ayurveda, Yunani and Chinese medicine etc. In this paper, the performance of existing techniques such as SVM classifier and KNN classifiers are compared with the proposed method. The metrics precision, recall, F-measure and accuracy are calculated and compared. The 100 images had taken for study and the proposed novel segmentation method gives the best accuracy. The experiment uses 480 (increase the dataset) images of eight types of abnormalities. 70% of images were used for training and 30% of images were used for testing. (Discuss the performance measure)