论文部分内容阅读
Metabolic information obtained by proton magnetic resonance spectroscopic imaging (1H-MRSI) has been approved to be a powerful tool to identify either benign or malignant glioma, as well as to confirm the tumor level. However, 1H-MRSI data are affected by various factors, such as the thermal noise, eddy currents, susceptibility artifacts, and rigid body motion. To get accurate quantitative metabolic information, the key problem is to assess the 1H-MRSI data quality. In this paper, we introduce a new evaluating system to filter the data, and a new method, called wavelet denoising method, to improve the data quality under the evaluating system. Experimental results on 1H-MRSI glioma data demonstrate that preprocessing is prerequisite and the proposed algorithm with evaluating system is effective.
Metabolic information obtained by proton magnetic resonance spectroscopic imaging (1H-MRSI) has been approved to be a powerful tool to identify either benign or malignant glioma, as well as to confirm the tumor level. However, 1H-MRSI data are affected by various factors , such as the thermal noise, eddy currents, susceptibility artifacts, and rigid body motion. To get accurate quantitative metabolic information, the key problem is to assess the 1H-MRSI data quality. In this paper, we introduce a new to system motion filter the data, and a new method, called wavelet denoising method, to improve the data quality under the evaluating system. Experimental results on 1H-MRSI glioma data demonstrate that preprocessing is prerequisite and the proposed algorithm with evaluating system is effective.