Evaluation of Multimodal Algorithms for the Segmentation of Multiparametric MRI Prostate Images

<div>Transverse views of the ADC and DWI images with the respective segmentations of PZ (red) and CG (blue) of ground truth, DenseVNet, HighRes3DNet, and ScaleNet for 4 subjects with (a) the highest WG’s DSC across all networks, (b) TURP, (c) abnormally large prostate volume with uneven image intensities, and (d) PUC on T2-weighted images.</div>

Computational and Mathematical Methods in Medicine

fig2

Figure 2

Figure 2: Evaluation of Multimodal Algorithms for the Segmentation of Multiparametric MRI Prostate Images 