Dice Coefficient Calculator
Calculate segmentation overlap quality from voxel counts or binary masks. Useful for validating medical image segmentation models and comparing manual vs AI masks.
Paste equal-length binary arrays using commas, spaces, or line breaks. The tool treats any non-zero value as foreground.
Dice score
overlap percent
Jaccard / IoU
non-overlap voxels
Formula
Dice = 2 × |A ∩ B| / (|A| + |B|). Jaccard/IoU = |A ∩ B| / (|A ∪ B|). Dice is commonly used to evaluate segmentation overlap, but it should be interpreted with anatomy size, class imbalance, boundary quality, and clinical context.
Building or deploying segmentation models?Use our Medical AI Deployment Checklist before moving from model to workflow.
Download Checklist