摘要
PROPELLER数据采集成像技术利用K空间中心重叠采样区域的数据来估计受检查者的运动并加以校正,能较好地消除运动伪影。本文提出了一种基于圆形网格化的PROPELLER旋转校正算法,将中心重叠区域数据网格化到圆形的网格点上,避免了旋转估计时需多次网格化的缺点;并提出有效的相似性测度公式,通过计算测度值估计相应的旋转运动参数,据此对各数据带进行旋转校正。实验表明,与传统旋转校正算法相比,该算法运行速度快,成像质量好。
PROPELLER data acquisition imaging technique can extract inter-strip motion information from data in central overlapped sampling area, and suppress the artifacts caused by patients＇ rigid motion effectively. This study proposed a cireular-gridding algorithm for rotation correction. It calculated data in central overlapped area onto circular grid points, and thus avoided much gridding computation. Besides, it proposed a new efficient similarity measure formula to calculate the value of measure, estimated the rotation parameters and corrected data strips. Experimental results demonstrated the proposed method was running faster than the conventional algorithm, and improved image quality.
出处
《北京生物医学工程》
2009年第5期477-480,共4页
Beijing Biomedical Engineering
