What is the difference between anchor-based and anchor-free object detectors?
Anchor-based detectors place a fixed set of reference boxes everywhere and predict adjustments to them. Anchor-free detectors drop those references and predict objects directly - usually as a center point plus size, or as keypoints. Anchor-free removes the dataset-specific anchor tuning, has fewer hyperparameters, and is what most recent YOLO versions use.
Anchor-based
Tile the image with reference boxes of set scales and aspect ratios, then predict an offset and a class for each. It works well but adds hyperparameters: how many anchors, what sizes, what ratios - and those often need re-tuning per dataset with k-means.
Anchor-free
Predict objects directly. Center-based methods (like FCOS) treat each feature-map location as a potential object center and regress the distances to the four box edges. No anchor boxes to design, fewer knobs, and often simpler training.
- Anchor-based: strong, mature, but anchor design is dataset-sensitive.
- Anchor-free: fewer hyperparameters, cleaner, now the default in modern YOLO and many current detectors.
For most people using an off-the-shelf modern YOLO, this is handled for you - you get the anchor-free benefits without touching anchors at all.
Related questions
Hands-on tutorials
Stuck on this in a real project?
Book a free call and I’ll give you a straight answer on your specific case.
Book a consultation