2000 character limit reached
Dice-Contour Regularization Loss
Updated 8 July 2026
- Dice-Contour Regularization Loss is a segmentation approach combining Dice loss with contour components to ensure precise boundary delineation.
- The method incorporates contour-aware regularization terms into the loss function to refine predictions by emphasizing edge precision.
- It has demonstrated improved segmentation performance on complex boundaries, outperforming traditional loss functions in various applications.
Searching arXiv for the cited works and closely related contour/boundary segmentation losses. I’m checking arXiv metadata for the most relevant contour- and boundary-aware segmentation losses, including Dice+Contour and boundary/surface formulations.