| Glioblastoma, the GBM, is the most severe form of gliomas and is the highest grade of gliomas.Magnetic resonance imaging(MRI) and Magnetic resonance spectroscopic imaging(MRSI) has become the most authoritative methods of detection in the medical field, especially in the detection of primary human brain tumors.However,neither of the two types of data can separate the tissues of GBM successfully,a fusion result with two kinds of data characteristics can be obtained by fusion method.The fusion of MRSI data and MRI data belong to multi-modal data fusion since MRSI data is spectral data and MRI data is the image data. Based on data fusion method,an unsupervised multi-modal data GBM tissue separation method is proposed using non-negative matrix factorization and wavelet theory. The main contribution is presented as follows:1?A multi-modal medical data fusion method based on wavelet decomposition is studied, which can be integrated into the hNMF(Hierarchical non-negative matrix factorization)framework for GBM recognition to improve the accuracy of recognition.2?A new rule is studied during hNMF to determine the characteristic spectrum,In view of the present situation of unstable metabolic characteristics of GBM cases,achieving a more effective factorization of MRSI.By determine the characteristic spectrum of normal, tumor and necrotic tissue, the spatial distribution of the corresponding tissues type was obtained.3?the MRSI data features were extracted by peak integration method,quantification of key metabolic components in MRSI data,through the experiment identified the extracted metabolites and the range of them, which still effective in separating the tissues and greatly reducing the iterative computation of NMF.4?A segmentation method based on FCM determine the specific boundaries of different types of tissues is studied,the final fusion results are obtained.Comparing the segmentation results and the expert labeling to prove the validity of the proposed framework. |
PDF Full Text Request