Difference between Precision and mAP50:95?

[yolor2]

Search before asking

• I have searched the YOLOv5 issues and discussions and found no similar questions.

Question

Could you please tell me the difference between prececision and mAP50:95 as while running sweeps, for some sweeps i am getting more precision like 84% and mAP50:95 around 54, while for some sweeeps both are relatively same. SO what could be the reason for that?

Additional

No response

[glenn-jocher]

See https://en.wikipedia.org/wiki/Precision_and_recall

[yolor2]

@glenn-jocher what do you infer if the precision is around 80% and the mAP50_95 is around 70 and mAP50 around 90

[glenn-jocher]

@yolor2 you can ignore P and R, they are relative metrics which depend on a confidence threshold. mAP is an absolute metrics that is not a function of confidence threshold.

[yolor2]

@yolor2 you can ignore P and R, they are relative metrics which depend on a confidence threshold. mAP is an absolute metrics that is not a function of confidence threshold.

@glenn-jocher If mAP is an absolute metrics then why change in confidence threshold changes the values of mAP? could you please help me this dilemma

[glenn-jocher]

@yolor2 mAP is not valid at any other value than the default confidence threshold or lower. Ideally it is only ever computed at 0.0 confidence by definition.

There is no correct mAP calculation at any other confidence.

[yolor2]

@glenn-jocher hello, what do you think is the best for object detection task : mAP50 or mAP50:95 and why? because i got mAP50 around 90% and mAP50:95 around 60%. So i am trying to find the justification of this gap.
Could you please help.

[glenn-jocher]

@yolor2 mAP@0.5 is the official VOC metric and mAP@0.5:0.95 is the official COCO metric. There is no 'best' metric, but mAP@0.5:0.95 is the most widely recognized object detection metric.

[github-actions]

👋 Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 🚀 resources:

• Wiki – https://github.com/ultralytics/yolov5/wiki
• Tutorials – https://docs.ultralytics.com/yolov5
• Docs – https://docs.ultralytics.com

Access additional Ultralytics ⚡ resources:

• Ultralytics HUB – https://ultralytics.com/hub
• Vision API – https://ultralytics.com/yolov5
• About Us – https://ultralytics.com/about
• Join Our Team – https://ultralytics.com/work
• Contact Us – https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 🚀 and Vision AI ⭐!