Deep Learning Combined with Radiologist’s Intervention Achieves Accurate Segmentation of Hepatocellular Carcinoma in Dual-Phase Magnetic Resonance Images

<div>A typical example of hepatocellular carcinoma (HCC) segmentation with high accuracy. The Dice similarity coefficient (DSC) was 0.86 and 0.92 for deep fusion network- (DFN-) F and DFN-R, respectively. All contouring results are presented on hepatobiliary phase- (HBP-) magnetic resonance imaging (MRI). (a) Ground truth (GT). (b) Segmentation result of deep convolutional neural network- (DCNN-) F. (c) Segmentation result of DFN-F. (d) Segmentation result of DFN-R.</div>

BioMed Research International

fig4

Figure 4

Figure 4: Deep Learning Combined with Radiologist’s Intervention Achieves Accurate Segmentation of Hepatocellular Carcinoma in Dual-Phase Magnetic Resonance Images 

Figure 4 | Deep Learning Combined with Radiologist’s Intervention Achieves Accurate Segmentation of Hepatocellular Carcinoma in Dual-Phase Magnetic Resonance Images 