Metrics: Mean Average Recall (mAR)

[LinasKo]

Metrics: Mean Average Recall (mAR)

Tip

Hacktoberfest is calling! Whether it's your first PR or your 50th, you’re helping shape the future of open source. Help us build the most reliable and user-friendly computer vision library out there! 🌱

We'd like to expand our suite of metrics with a new one - Mean Average Recall (mAR). This would involve creating it, its accompanying results class, and briefly testing it.

Note it is different from mAP in that it don't use the precision-recall curve, but recall-IoU. This affects MeanAverageRecallResult:

• The will not be mAR@50, mAR@75, etc - only the global main mAR value.
• Other frameworks threshold by max detections. I suggest we add this later if required.
• mAP currently implements a single averaging method, and F1 has 3 different ones (micro, macro, weighted). The choice between those is left to PR author - we can always add more later.
• The metric can report mAR values per-class. Do tell if that is difficult.
• I think we should have 1.0 as the default if no value are provided. -1 is a consideration too, which should bring the same change to mAP.

Feel free to change the above if it feels better.

---

Helpful links:

• Contribution guide
• Metrics:
  • mAP metric: docs, code
  • F1 Score: docs, code
• Supervision Cheatsheet
• Colab Starter Template
• Prior metrics test Colab

[LinasKo]

Contribution guidelines

If you would like to make a contribution, please check that no one else is assigned already. Then leave a comment such as "Hi, I would like to work on this issue". We're happy to answer any questions about the task even if you choose not to contribute.

Testing

Please share a Google Colab with minimal code to test the new feature. We know it's additional work, but it will speed up the review process. You may use the Starter Template. The reviewer must test each change. Setting up a local environment to do this is time-consuming. Please ensure that Google Colab can be accessed without any issues (make it public). Thank you! 🙏

[LinasKo]

@onuralpszr, I've heard you already put some work towards this one. Shall I assign it to you?

[onuralpszr]

@onuralpszr, I've heard you already put some work towards this one. Shall I assign it to you?

Thank you and yes

[LinasKo]

Hey @onuralpszr, how's it going? Did you manage to get the mAR working in the end?

[onuralpszr]

Hey @onuralpszr, how's it going? Did you manage to get the mAR working in the end?

After my work let me check and open pr

[LinasKo]

Implemented via #1661