harryhan618 / LaneNet

Pytorch implementation of "Towards end-to-end lane detection: an instance segmentation approach"
166 stars 42 forks source link

about quantitative evaluation #11

Closed somebodyus closed 4 years ago

somebodyus commented 5 years ago

When will you publish your test code? I used the postprocess code from MaybeShewill-CV/lanenet-lane-detection, and I got very low accuracy.

harryhan618 commented 5 years ago

Hello! What kind of metric do you want for evaluation? mIOU would be easy to implement. So I would update it these days. CULane evaluation is a little bit tricky since this is a binary sementation method. So I would be happy to discuss about it.

somebodyus commented 5 years ago

I mean the Accuracy, FP, FN that are defined in the LaneNet paper: Towards end-to-end lane detection: an instance segmentation approach