Skip to main content
Fig. 1 | Radiation Oncology

Fig. 1

From: A dual deep neural network for auto-delineation in cervical cancer radiotherapy with clinical validation

Fig. 1

Clinical target volume (CTV) contouring predicted by SegNet(A, B, C). The orange, turquoise, purple, and red contours represent the CTV segmented by SegNet(A), the CTV by SegNet(B), the CTV by SegNet(C), and the corresponding manual annotations, respectively. The first and second columns indicate the areas of the common iliac lymph area and the internal and external iliac lymph nodes, respectively. The third and fourth columns present the presacral lymph nodes area and the end of this area, respectively. The fifth and sixth columns indicate the paravaginal tissue and upper vagina areas, respectively. In Case a, the volumetric DSC quantitative scores by SegNet(A), SegNet(B), and SegNet(C) were 0.81, 0.86, and 0.82, respectively. In cases b–d, the volumetric DSC scores for SegNet(A), SegNet(B), and SegNet(C) were 0.87/0.84/0.81, 0.80/0.86/0.85, and 0.80/0.86/0.82, respectively

Back to article page