Dynamic Regulation of Level Set Parameters Using 3D Convolutional Neural Network for Liver Tumor Segmentation

Figure 11

Segmentation results obtained by the proposed method on the 3DIRCADb dataset. (a)–(p) One case of tumor segmentation results in adjacent slices and the 3D reconstruction result. (q) 3D reconstruction result of the liver and its entire tumors are inside. Green is the ground truth annotated by radiologists. Red is the segmentation result by DRLS.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

(k)

(l)

(m)

(n)

(o)

(p)

(q)