Figure 2 | Scientific Reports

Figure 2

From: Deep learning for fully automatic detection, segmentation, and Gleason grade estimation of prostate cancer in multiparametric magnetic resonance images

Figure 2

Output of the model (every row corresponds to a different patient) evaluated on two IVO test patients (first two rows) and three ProstateX test patients (last three rows). For each patient, first image from the left shows the ground truth on the T2 sequence; the rest show the output predictions of the model on different sequences (from left to right: T2, b800, ADC, \(K^{trans}\) -IVO- / DCE \(t=30\) -ProstateX-). Gleason Grade Group (GGG) 0 -benign- bounding boxes (BBs) are not shown and only the highest-scoring BB is shown for sets of highly overlapped detections (intersection over union \(>0.25\)). Detections with confidence below the GGG\(\ge\)2 lesion-wise maximum sensitivity threshold (0.173 for IVO, and 0.028 for ProstateX) are not shown either.

Back to article page