Mean Average Precision (MAP)
Definition
MAP averages Average Precision values according to a defined evaluation protocol.
MAP averages Average Precision in information retrieval over queries; in object detection, averaging is done depending on the protocol, e.g., over classes and IoU thresholds.
Simply explained
How to evaluate multiple rankings individually and then aggregate their results; the rules must be consistent.
Why it matters in practice
Helps realistically assess model quality, limits, and technical decisions.
Example from business
A detection model is evaluated over multiple object classes and IoU thresholds.
Limits and pitfalls
- Do not directly compare MAP values from different protocols; in object detection, also specify IoU thresholds.
Origin & history
mAP was popularized through information retrieval and object detection benchmarks; VOC and COCO use different variants.
Use in mid-sized companies
- Possible application: A team defines the evaluation protocol and test data for a search or detection task. Results are only compared under identical averaging rules and conditions.
Frequently asked questions
What does Mean Average Precision (MAP) mean?
MAP averages Average Precision in information retrieval over queries; in object detection, averaging is done depending on the protocol, e.g., over classes and IoU thresholds.
What should one pay attention to?
Do not directly compare MAP values from different protocols; in object detection, also specify IoU thresholds.
Sources
- cocodataset.org
- Ergänzender Primärbeleg · Ergänzender Primärbeleg
- Ergänzender Primärbeleg · Ergänzender Primärbeleg
Source verified on 10.10.2026. Editorial application examples are hypothetical; no evidence of real customer results. Additional sources verified on 12.10.2026.
Editorially reviewed on: