| 摘要 |
A method acquires three complex images, wherein their respective phase differences between water and fat components are alpha 0, alpha 0+ alpha , and alpha 0+2 alpha , respectively. The method obtains from these three complex images two possible solution sets for water and fat images. For pixels with both water and fat components, one correct solution set is selected using a binary choice based on the relative Larmor frequencies of water and fat. If a pixel contains only one component, a known statistical bias is applied to identify the component and, thus, the pixel. To correct misidentified pixels, various filters may be applied to all the pixels. The water and fat solutions obtained from the three complex images are used to produce separate images of water and fat. Second pass solutions of water and fat with improved signal-to-noise ratio can be obtained by either averaging or a least square error method.
|
