基于LBP块相似性的图谱标签融合

Atlas Label Fusion Based on the LBP Block Similarity

磁共振脑部图像的定量分析通常需要对脑部解剖结构进行精确可靠的自动分割,基于多图谱的脑部核磁共振(MRI)图像自动分割方法是解决该问题的一种非常有效的方法,其中比较关键的一个问题是图谱标签融合问题.本文针对该问题提出了一个基于局部二进制模式(LBP)块相似性的图谱标签融合方法.实验结果表明,利用该方法对脑部MRI图像中的组织进行分割,能够获得较精确的分割结果.

Quantitative MRI analysis often requires accurate and reliable automatic segmentation of brain anatomical structures, and the automatic segmentation based on multi-Atlas brain MRI image is an effective method. During the process, atlas label fusion is the key issue. In this paper, an atlas label fusion method based on LBP block similarity is proposed. The paper demonstrates that the proposed method can obtain a more accurate segmentation result of the brain MRI image.