Multiple kernel learning with adaptive kernel method for computer-aided detection of colonic polyps

Computer-aided detection (CAD) of colonic polyps, as a second reader for computed tomographic colonography (CTC) screening, has earned extensive research interest over the past decades. False positive (FP) reduction in the CAD system plays a crucial role in detecting the polyps. To improve the performance of FP reduction and better assist the physician's diagnosis, we propose a multiple kernel learning (MKL) with adaptive kernel method for CAD of colonic polyps, called AK-MKL method. Using the multiple kernel learning technique, the AK-MKL method learns a synthesized classifier which is an optimal combination of a collection of base classifiers. Performance evaluation for the presented AK-MKL method was performed on a CTC database. In terms of the AUC (area under the curve of receiver operating characteristic) merit, the experimental results showed that our AK-MKL method achieves better performance, compared with other two different methods, named the basic multiple kernel learning method (MKL) and the SVM with adaptive kernel (AK-SVM) method, respectively.