Figure 6 | Scientific Reports

Figure 6

From: Self-supervised representation learning using feature pyramid siamese networks for colorectal polyp detection

Figure 6

Generalizability of the FPSiam method to another dataset under various challenging conditions. Four representative hard cases in the CVC-VideoClinicDB dataset are used to evaluate FPSiam’s robustness and adaptability. The green bounding boxes are the ground truth of polyps in the frames. Blue bounding boxes and red bounding boxes are the predictions of the FPSiam method and TL method, respectively.

Back to article page