Biomedical subjects
Publications and source records attributed to Shan'an Zhu.
Medical image segmentation is an important application of image segmentation. However it is the bottleneck that restrains medical image application in clinical practice. In this paper, the aim and significance of medical image segmentation are discussed, the development of medical image segmentation techniques is sketched, and a review of the medical image segmentation techniques is given.
We have investigated spatio-temporal source modeling (STSM) of the electroencephalogram (EEG) by using a Quasi-Newton method based on Independent Component Analysis (ICA) for localization of multiple dipole sources from the scalp EEG. The problem of multiple dipole localization was transformed into several single dipole localization problems. Another benefit of the present method is that the number of independent sources can be estimated. Computer simulation studies were conducted to evaluate the performance of this approach. The present simulation results indicate that the ICA-based method is superior to the conventional nonlinear methods in localization accuracy, computation time and anti-noise performance, for multiple dipole localization when the sources are stationary over the period of interest.