![]() ![]() A second factor (especially for higher field strength) is the occurrence of spatially dependent excitation flip angles, leading to non-uniform contrast weighting in MR images. Especially when using surface coils or coil arrays for recording sensitive data, one finds clearly variable grey values in the same image for a given type of tissue, depending on the distance of the tissue site from the receiver coils. First, grey values in MR images are proportional to signal intensities recorded by the receiver coils rather than the signal intensity generated inside the tissue. On the other hand, some characteristics of MR images create greater difficulty for automatic tissue segmentation than CT data sets. On the one hand, MRI, with its multiple forms of contrast weighting and its sensitivity to diffusion, perfusion, and chemical composition of tissues, is a highly attractive alternative. MRI is clearly the preferred imaging modality for most cohort studies with healthy volunteers and for studies requiring optimal soft tissue contrast. On the other hand, the recording of highly resolved data sets by CT is problematic because of inevitable radiation exposure (especially for multiple scanning sessions in functional studies), and soft tissue contrast of CT data is inherently inferior to that of MRI. Related post-processing procedures are available on modern CT consoles working fully automatically. Angiographic CT data sets (including coronary angiography) recorded after intravenous administration of contrast media can be successfully reconstructed using maximum-intensity projection techniques (MIPs), but higher-quality and better visualization of the spatial relation between blood vessels and other body parts is achieved by surface reconstruction after tissue segmentation. Thus, picture elements in CT scans can be attributed to several tissue classes (e.g., soft tissue or bone) using relatively simple threshold approaches. CT based on X-ray absorption provides a favorable basis for automatic segmentation of several tissue classes, as grey values are given in well-defined Hounsfield units that directly reflect X-ray absorption in the tissue. Some structures, however, such as the faces of fetuses, can often be reconstructed well. PET alone (without congruent CT or MR images) provides relatively low spatial resolution, and US images often show strong artefacts that may not be adequately handled by automatic segmentation software. Common imaging modalities useful for recording tomograms or isotropic three-dimensional data in human examinations include computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET), and ultrasound (US). ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |