Automate Image Analyses for Magnetic Resonance Imaging
Abstract (Set) In this disclosure, we describe methods to perform automated analyses of MRI data. We propose automated tissue segmentation and subsequent morphological and photometrical analyses. The main goal is computer assisted diagnosis (CAD) but the technique can also be applied to a wide range of MRbased researches. The proposed method consists of eight components: 1) Smar image acquisition, 2) Smart preprocessing, 3) Automated tissue segmentation, 4) Automated morphological analysis, 5) Automated photometric analysis, 6) Image database, 7) Data presentation, and 8) Clinical database. Each component can be used as an effective tool for CAD, but by combining all eight components, the efficiency and accuracy of CAD increases. Description (Set) There are four important components in this package. These are; 1) intensity homogeneity corrections, 2) multi-contrasts, 3) skull-suppressed contrast, and 4) quantitative MRI. 1) Intensity homogeneity corrections: This technique is already widely used and becoming available to many clinical scanners. This ensures that the intensity profile of the imaged object is homogeneous, which facilitates the efficiency and accuracy of subsequent image analysis. 2) Multi-contrasts: To use our CAD technique, it is preferable to employ multi-contrasts such as T1, T2, inversion-recovery prepulse, diffusion, perfusion, and diffusion tensor. Each contrast delineates different brain structures better than others. For example, T1 provides high contrast to differentiate the gray and white matter, while T2 is suitable to define the ventricle. It is also preferable that the CAD integrates detailed imaging parameters such as echo time, repetition time, inversion recovery time, field of view and imaging matrix. This ensures homogeneous quality of analyzed images. 3) Skull-suppressed contrast: One of the largest obstacles of automated brain extraction is contamination by adjacent tissues such as dura, fat, and skin. Signal intensity of MR images is influenced by various physical and chemical properties of water molecules such as T1, T2, T2*, proton density (PD), and diffusion (D) constant. 4) Quantitative MRI Images obtained from MRI has arbitrary signal intensity as shown in equation 1, which is a function of various parameters. However, by acquiring more than one image with different imaging conditions, we can quantify such parameters as T1, T2, and D. This type of quantification approaches can facilitate comparison between two images from different subjects or different scanners. Proposed Use (Set) This routine can be incorporated into commercial MRI scanner or off-line workstation for MRI image analysis.Inventor(s): Mori, Susumu ,Miller, Michael,Zhang, Jiangyang
Type of Offer: Licensing
« More Medical Patents