Deep Learning-Based Acute Ischemic Stroke Lesion Segmentation Method on Multimodal MR Images Using a Few Fully Labeled Subjects

Figure 4

Visualization examples of the MRI slices and lesion segmentation results. (a–c) The original ADC map, DWI, and ground truth, respectively. (d–h) The segmentation results of U-Net, FCN-8s, Res-UNet, the method in [31], and the proposed method, respectively. The segmentation results are overlaid on the DWIs and highlighted in red.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)