Affiliations: [a] School of Earth Sciences and Engineering, Sun Yat-Sen University, Guangzhou 510275, China. [email protected] | [b] Department of Neurology, First Affiliated Hospital of Nanjing Medical Univ., Nanjing 210029, China. [email protected] | [c] School of Architecture Building and Civil Engineering, Loughborough University, Loughborough, LE11 3TU, UK. [email protected] | [d] School of Earth Sciences and Engineering, Sun Yat-Sen University, Guangzhou 510275, China. [email protected]
Correspondence:
[†]
Address for correspondence: School of Architecture Building and Civil engineering, Loughborough University, Loughborough, LE11 3TU, UK and also: School of Earth Sciences and Engineering, Sun Yat-Sen University, China
Note: [*] Those authors contributed equally to this paper
Abstract: The existence and distribution pattern of cerebral microbleeds (CMBs) are associated with some underlying aetiologies caused by intra-cerebral hemorrhage (ICH). CMBs as a kind of subclinical sign can be recognized via magnetic resonance (MR) imaging technique in a few years before the onset of the disease. Hence, detecting CMBs accurately is important for treating and preventing related cerebral disease. In this study, we employed convolution neural network (CNN) for CMBs detection because of its powerful ability in image recognition. In view of too many efforts on optimizing the structure of CNN for achieving a better performance, we introduced center loss, which can greatly enhance the discriminative power of the deeply learned features, to CMBs detection for the first time. It is found that the performances of convolution neural network (CNN) trained under the joint supervision of softmax loss and center loss were significantly better than that under the supervision of softmax loss, even if there are few mislabelled samples in training data. With this trick, we achieved a high performance with a sensitivity of 98.869 ± 1.026%, a specificity of 96.491 ± 0.367%, and an accuracy of 97.681 ± 0.497%, which is better than four state-of-the-art methods.
Keywords: cerebral microbleeds, convolution neural network, discriminative feature learning, center loss
