search

lucastabelini / PolyLaneNet

Code for the paper entitled "PolyLaneNet: Lane Estimation via Deep Polynomial Regression" (ICPR 2020)
https://arxiv.org/abs/2004.10924
MIT License
295 stars 74 forks source link

regarding the dataloader #38

Closed seyeeet closed 3 years ago

seyeeet commented 3 years ago

can you please explain the strategy behind this line of code? link lanes = np.ones((self.dataset.max_lanes, 1 + 2 + 2 * self.dataset.max_points), dtype=np.float32) * -1e5 what/why we have 1, and 2,2* self.dataset.max_points? why do we multiply it with * -1e5?

lucastabelini commented 3 years ago

1 (confidence score) + 2 (bottom and top vertical positions) + 2 * max_points (a X and Y position for each point) The array is multiplied with -1e5 just to make bug detection easier, it's just an in implementation detail. You can ignore it. In case the value is used somewhere it would be easier for me to debug. It is also used here.

seyeeet commented 3 years ago

Thank you for the explanation! can you also tell me what category represent in the dataloader?

lucastabelini commented 3 years ago

It's the category of lane (e.g., dotted, dashed, etc). The only dataset I used that provides this information is ELAS. In the other datasets, the two categories are just positive (the lane exists) and negative (it doesn't).