基于局部位移校正的磁共振图像相干平均

发布时间：2018-02-24 11:34

本文关键词： 磁共振成像(MRI) 相干平均非局域均值图像去噪　出处：《波谱学杂志》2017年03期 　论文类型：期刊论文

【摘要】：多次扫描相干平均是提高磁共振图像信噪比的常用方法,但如果在多次扫描过程中病人发生自主或不自主的运动,使得图像中的组织发生位移,简单相干平均图像会导致图像模糊.本文受非局域均值算法的启发,提出了一种基于局部位移校正的相干平均方法.该算法通过比较多次采集的图像中组织结构的局部相似性,找出图像间的局部位移,利用该信息修正位移后进行加权平均,从而达到提高图像信噪比的目的.我们用模型及真实的肝脏弥散数据进行了实验.实验结果表明,对于不同次采样间存在运动的磁共振图像,该算法可有效地提高信噪比并保持结构边缘;其结果优于简单的相干平均,去噪效果也优于经典的非局域均值算法.
[Abstract]:Multi-scan coherence averaging is a common method to improve the SNR of magnetic resonance image. However, if the patient moves autonomously or involuntarily in the process of multiple scans, the tissue in the image will shift. A coherent averaging method based on local displacement correction is proposed in this paper, which is inspired by the nonlocal mean algorithm. We find out the local displacement between images and use the information to correct the displacement and carry out the weighted average, so as to improve the SNR of the image. We use the model and the real liver diffusion data to carry out the experiment. The experimental results show that, For the magnetic resonance images with motion between different subsamples, the proposed algorithm can effectively improve the SNR and keep the structure edge, and the result is better than the simple coherent average, and the denoising effect is better than the classical nonlocal mean algorithm.
【作者单位】：华东师范大学物理与材料科学学院上海市磁共振重点实验室;上海卡勒幅磁共振技术有限公司;西门子医疗东北亚科研合作部;上海健康医学院;广州市番禺区中心医院放射科;
【基金】：国家高技术研究发展计划(“863”计划)资助项目(2014AA123400)
【分类号】：R445.2;TP391.41

【相似文献】