奇异值分解滤波联合Frangi滤波对超微血管成像图像质量的影响

目的 观察奇异值分解(SVD)滤波联合Frangi滤波对超微血管成像(SMI)图像质量的影响。方法 分别对1具仿体及1例疑诊肝血管瘤患者行CEUS,分别以SVD滤波(A组)、Frangi滤波(B组)、Frangi+SVD滤波(C组)及SVD+Frangi滤波(D组)处理图像,之后比较不同图像的质量。结果 4组仿体及人体肝脏SMI对比组织比(CTR)、对比度噪声比(CNR)及信噪比(SNR)均高于原始CEUS图像。相比原始图像,D组图像质量及分辨率提升最为显著,仿体SMI的 CTR、CNR及SNR分别提升58.04、3.39及48.04 dB,人体肝脏SMI相应参数则分别提升61.85、16.80及49.67 dB,而分辨率分别为原始CEUS图像的1.42倍及1.98倍。结论 SVD滤波联合Frangi滤波可有效提高SMI图像质量。

Impact of singular value decomposition filter combined with Frangi filter on image quality of superb microvascular imaging

Objective To observe the impact of singular value decomposition (SVD) filter combined with Frangi filter on image quality of superb microvascular imaging (SMI). Methods Contrast-enhanced ultrasound (CEUS) was performed in a phantom and a patient with suspected liver hemangioma. Images in group A were processed with SVD filter, while in group B, C and D with Frangi filter, Frangi+SVD filter and SVD+Frangi filter, respectively, then image quality of SMI were compared. Results Contrast-to-tissue ratio (CTR), contrast-to-noise ratio (CNR) and signal-to-noise ratio (SNR) of A,B, C and D group were all higher than those of original CEUS images. Compared with the original CEUS images, the image quality indexes and image resolution of group D were most significantly improved, with CTR, CNR and SNR increased by 58.04, 3.39 and 48.04 dB of the phantom, as well as increased by 61.85, 16.80 and 49.67 dB of the patient, respectively. The resolution of the phantom''s and the patient''s SMI images was 1.42 times and 1.98 times of the original CEUS images, respectively. Conclusion SVD filter combined with Frangi filter could effectively improve image quality of SMI.