采用自动选择参照图校正肝超声造影图像呼吸运动

参照图选择是肝超声造影呼吸运动校正方法有效性的关键因素之一,其选择方法值得进一步探索。先采用拉普拉斯特征映射(LE)法将原始高维超声数据降至二维空间,再采用k-means法进行聚类分析,获取理想的超声参照图进行呼吸运动校正,最后在18例兔肝VX2肿瘤超声造影病例上验证并评价该方法的有效性。校正前图像序列的平均结构相似度和平均互相关系数均值分别为0.45±0.11和0.67±0.16,校正后两者均明显增加(P<0.001),分别为0.59±0.11和0.81±0.11。校正前图像序列的偏差值(DV)均值为92.16±18.12,校正后均值下降到约为原值的1/3。结果表明,所提出的呼吸运动校正方法能提高传统依靠人工选择参照面进行超声定量分析方法的准确性,且操作简单,具有一定临床应用潜力。

Respiratory Motion Correction of Liver Contrast-Enhanced Ultrasound Sequences by Selecting Reference Image Automatically

The selection of reference image is one of the key roles for the efficiency of respiratory motion correction method based on liver contrast-enhance ultrasound (CEUS) image sequences. The selection method is worthwhile exploring. First, the original high-dimensional ultrasound data was mapped into a two-dimensional space by using Laplacian eigenmaps. Then, the cluster analysis was adopted using k-means, and the optimal ultrasound reference image was obtained for the respiratory motion correction. Finally, the effectiveness of this proposed method was validated with 18 CEUS cases of VX2 tumor implanted in the rabbit liver. Before correction, the average of total mean structural similarity and the average of mean correlation coefficient from image sequences were 0.45±0.11 and 0.67±0.16,respectively. After correction, the two parameters were increased obviously (P<0.001) as 0.59±0.11 and 0.81±0.11 respectively. The average of deviation valve (DV) from image sequences before correction was 92.16±18.12. After correction, the average of DV was reduced to one-third of the original value. The proposed respiratory motion method improved the accuracy of the quantitative analysis of CEUS by using the reference image based on the traditionally manual selection as well as operated simply, therefore is of potentials in the clinical application.