F-score feature selection method may improve texture-based liver segmentation strategies

A fast computer-aided liver segmentation plays a vital role in computer aided surgery (CAS), especially when using texture-based methods. Large amount of features yielded in supervised segmentation methods may result in slow segment processes. In order to reduce the time required in the segment stage, we applied principal component analysis (PCA), forward orthogonal search by maximizing the overall dependency (FOSMOD), and F-score to our supervised method proposed recently. Our results showed that the F-score may help in accelerating segment stage by approximately 42% whilst the PCA-based feature selection method failed to extract the liver contour correctly. Though FOSMOD can obtain a good segmentation result of liver, it is time consuming comparing with the other two methods. Thus, F-score method may provide an effective solution for accelerating liver segmentation.