Command Palette
Search for a command to run...
Boundary F1
The Boundary F1 Score is a metric specifically used in computer vision image segmentation tasks to evaluate boundary quality. It computes the harmonic mean of pixel-level precision and recall between the boundaries of predicted segmentation masks and ground-truth annotation boundaries, measuring contour alignment. It is particularly sensitive to thin or elongated structures, ambiguous boundaries, or small objects. The value is expressed as a percentage, typically reported as a percentage or a 0–1 score; higher values are better, indicating that predicted boundaries are closer to ground-truth boundaries. It is often used together with region-overlap metrics such as mIoU and Dice to supplement segmentation quality evaluation from the boundary dimension.
Build AI with AI
From idea to launch — accelerate your AI development with free AI co-coding, out-of-the-box environment and best price of GPUs.