Things we can try from solutions

[mrinal18]

@nvk681 @lalalaurenville @gsvsumanth @Archana-bc

1. Confidence : 0.01 - 0.7
2. IOU score : 0.01 or 0.35
3. Angle (havent tried yet)
4. Direction of moment (havent tried yet)
5. Epochs < (10-15)
6. Tracking of helmets
7. Limited players <= 22
8. Check for outlier bounding boxes by taking fps as consideration. : needs experimenting.

What we have now
9. Baseline model: 0.6
10. Tracking angles : 0.806

Something to learn:
Hungarian Algorithm

What we can do: Next Wednesday

1. https://www.kaggle.com/c/nfl-health-and-safety-helmet-assignment/discussion/285286

2. https://www.kaggle.com/c/nfl-health-and-safety-helmet-assignment/discussion/285076

3. https://www.kaggle.com/c/nfl-health-and-safety-helmet-assignment/discussion/285065

4. https://www.kaggle.com/c/nfl-health-and-safety-helmet-assignment/discussion/284940

5. Try with different values IOU, conf
   Tracking angles
   -> Check for outlier bounding boxes by taking fps as consideration. : needs experimenting. People can split and try to experimenting
   -> Tracking of helmets

Notebooks to work on:

1. https://www.kaggle.com/mrinalmathur/testing-deepsort/edit/run/77737624 : deep sort + yolo5 only

2. https://www.kaggle.com/mrinalmathur/deepsort-with-tracking-angles?scriptVersionId=78732225 deepsort + yolo5 + tracking