Analysis of LBP and LOOP Based Textural Feature Extraction for the Classification of CT Lung Images

Lung Cancer tops the list among all cancers. According to a study by IASLC (International Association For the study of Lung Cancer) it is found that more than 1.6 million deaths are witnessed every year due to Lung Cancer, which is more than the death rate caused by prostrate, colon and breast cancers combined. Thus there is a need for an early detection followed by early treatment in order to improve the patient's chance of survival. In this paper a Lung Cancer detection model is developed using image processing technique. This model involves three stages to detect the presence of cancer nodule which are preprocessing, feature extraction and classification. The extracted features classify the lung as normal or abnormal with the help of SVM classifier. In this paper we extract texture features using Local Optimal Oriented Pattern(LOOP) and classify them using K-fold cross validation technique. The results obtained are then compared to the results of various binary patterns-LBP(Local Binary Pattern),LBC(Local Binary Count) and LDP(Local Directional Pattern).