| Summary: | 碩士 === 國立中興大學 === 電機工程學系 === 92 === Magnetic resonance imaging(MRI)is a very versatile medical imaging tool.The main reasons are twofold.First,It can produce high－quality images.Second,it has many imaging parameters that can be changed to fit different imaging needs.Because MRI requires a long process of exciting and receiving the nuclear magnetic resonance signal, the main hurdle for MRI today is its slow imaging speed. To overcome the problem,many fast imaging methods have been invented.Spiral MRI is one of these methods. Spiral imaging uses a spiral trajectory to quickly cover a large portion of k-space so that the number of times required to excite the NMR signal is reduced.Therefore,It can finish much faster than other methods.Comparing to other fast imaging methods,the benefit of the spiral imaging is that it is insensitive to motion-related artifacts. The problem is that spiral trajectory does not acquire data at rectilinear grid and it is necessary to re-grid the acquired data before fast Fourier transform. This process is often also called "gridding".
This thesis describes an implementation of the gridding algorithm and inverse fourier transform using JAVA RMI on multiple platfomrs.
The system architecture,we used is to pass the reconstruction parameters to RMI SERVER for gridding computation and transmit reconstructed data back to a personal computer for post processing.We validate that the system can successfully reconstruct spiral images distributedly.
|
|---|
