Performance Analysis of Certain Classifiers for Liver CT Images

Liver is the largest internal organ and is vital for the human body’s survival. It is prone to many diseases such as Liver tumor, Fibrosis, etc. In order to know the condition of the liver, the most commonly preferred method is biopsy. But since it is a very complex and complicated procedure it is being replaced by computer-aided diagnosis (CAD) where the liver is classified into various types like normal, abnormal etc. In this study, we have compared the performance of the CAD systems namely six classifiers for CT image classification. The data of 26 patients was taken into consideration and their status was confirmed by a radiologist. The images were separated into normal and abnormal based on textural features and based on these features the performance of each classifier has been evaluated for the parameters such as accuracy, specificity and sensitivity. Amongst all the classifiers we found out that the best results were obtained for k-NN with accuracy of 88.5%