Closed zhangmanCN closed 8 months ago
Please refer to the code about how to calculate P() in OpenLane-V2: https://github.com/OpenDriveLab/OpenLane-V2/blob/master/openlanev2/centerline/evaluation/evaluate.py#L331-L335
@zhangmanCN Sorry, we indeed made a mistake in the last row: 0.25 and 0.3. It should be 0.6 and 0.6. I will correct our manuscript.
Thank you for your reply. I think the 0.6 and 0.6 is right~
May I ask why the fifth result of prediction calculation before and after enhance prediction is 0.25 and 0.3 respectively.
In my understanding, as shown in the follow formula(it is a part of formula about TOP_ll):
To the ith predict topology of a certain vertex, and if it is correct, the predict topology score should be j/i * 1, where j is the right number in the first i topology predict result. and j/i is the result of the P(),1 is the result of the indicator func.
Do I have any mistake? Please advise me.