Skip to main content
Fig. 8 | BioMedical Engineering OnLine

Fig. 8

From: A robust statistics driven volume-scalable active contour for segmenting anatomical structures in volumetric medical images with complex conditions

Fig. 8

Comparison of our method with ITK-SNAP and Seg3D on five real brain tumor volumetric MR images. Column 1 the initial labels (red bubbles) and slices of the original volumetric images with complex image conditions; column 2 ground truth; column 3 results of ITK-SNAP for five images with region competition force α and the smoothing force β represented as a pair (α, β) = (0.65, 0.35), (0.65, 0.35), (0.55, 0.25), (0.5, 0.03), and (0.7, 0.25) in the order from left to right; column 4 results of Seg3D with threshold range TR, curvature weight CW, and propagation weight PW represented as a triple (TR, CW, PW) = (2,5, 0.1), (3.8, 7, 0.1), (2.55, 3, 0.1), (2.3, 3, 0.1), and (2.1, 3, 0.1) in the order from left to right; column 5 results of the proposed method

Back to article page