Automated Segmentation of Colorectal Tumor in 3D MRI Using 3D Multiscale Densely Connected Convolutional Neural Network

Figure 4

Qualitative comparison of colorectal tumor segmentation results produced by each method. In (a), from left to right columns are the raw MRI input volume and cropped volume, first three columns correspond to predicted probability by 3D FCNNs, and segmentation results by 3D FCNNs (red) and 3D FCNNs + 3D level set (red) overlapped with true ground truth (green), correspondingly. Similarly, second, third, and fourth three columns are related to predicted probability and segmentation results by rest of methods: 3D U-net (red), 3D U-net + 3D level set (red), DenseVoxNet (red), DenseVoxNet + 3D level set (red), 3D MSDensenet (red), and 3D MSDensenet + 3D level set (red), respectively. In (b), we have overlapped the 3D masks segmented by each method with the ground truth 3D mask. In (b), from left to right are ground truth 3D mask, overlapping of segmented 3D mask by 3D FCNNs (red), 3D FCNNs + 3D level set (red), 3D U-net (red), 3D U-net + 3D level set (red), DenseVoxNet (red), DenseVoxNet + 3D level set (red), 3D MSDensenet (red), and 3D MSDensenet + 3D level set (red) with the ground truth 3D mask (green points). The green points which are not covered by the segmentation results (red) of each method are referred as false negatives.

(a)

(b)