What question did this study set out to answer?

This study aims to improve pancreas segmentation accuracy in CT images using a two-stage pipeline.

June 10, 2026Applied SciencesOpen Access

Pancreas Segmentation Using a Two-Stage Pipeline of Faster R-CNN and TransUNet

View Full Paper

Ask AI

Bookmark

This study aims to improve pancreas segmentation accuracy in CT images using a two-stage pipeline.
Developed a two-stage segmentation framework combining Faster R-CNN and TransUNet.
Utilized a balanced training set from the NIH pancreas CT dataset with downsampling techniques.
Implemented a novel DHD loss function to improve segmentation accuracy at both region and boundary levels.
Achieved a mean DSC of 88.98%, precision of 91.0%, and recall of 94.4% on the NIH dataset.
Obtained a mean DSC of 62.96% in a zero-shot cross-dataset setting using the BTCV dataset.
Evaluated a fully automatic pipeline yielding a mean DSC of 66.50% ± 8.55% in 5-fold cross-validation.

No takes yet. Share an insight, caveat, or question.

Randomized trial demonstrates improved pancreas segmentation in CT images, suggesting enhanced accuracy.

Hong et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark