Segmentation
Boundary IoU
Implemented
segmentation.boundary_iouDefinition
IoU computed on contour bands of fixed width, emphasising boundary quality.
Formula
Boundary IoU = |(G_d ∩ G) ∩ (P_d ∩ P)| / |(G_d ∩ G) ∪ (P_d ∩ P)|
Range: [0, 1]
Inputs and outputs
- y_true: ground-truth mask (H×W or N×H×W)
- y_pred: predicted mask with the same shape
- dilation ratio d
Returns: float
Assumptions
No assumptions beyond valid, aligned inputs of the documented types.
Limitations
No metric-specific limitations are documented yet. Interpret the value alongside the task, data, and other metrics.
Python API
import evalsuite as es
es.boundary_iou(y_true, y_pred, dilation_ratio=0.02)References
- Cheng, B., Girshick, R., Dollár, P., Berg, A. C., & Kirillov, A. (2021). Boundary IoU: Improving object-centric image segmentation evaluation. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 15334–15342.