Computed Tomography Liver Tumor Image Classification Using Hybrid Feature Extraction and Classification Techniques

Accurate diagnosis of liver tumor is important task for the physician. The purpose of this research work is to investigate the utility of feature extraction and classification techniques in the diagnosis of liver tumor CT images as benign(non-cancerous) and malign(cancerous). The classification results can give useful information for the physician in diagnosing the tumor. In this work, the important features representing the tumor are extracted using SURF, LTP, and PCA. Then extracted features are passed to LSTM to identify the type of liver tumor. Experiments are carried out on the dataset collected from hospital. It is observed that the proposed method performs better when compared to the radiologist results.