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

What if we do not predict s and h #56

Closed Jensen-Su closed 2 years ago

Jensen-Su commented 2 years ago

I wonder if the model can fit a polyline x = a * y**3 + b * y ** 2 + c * y + d without fitting offset s and horizontal line h, since the problem seems to be poor-formed.

Intuitively, at point y = 100pix, the equation would be x = a * 1000000 + b * 10000 + c * 100 + d. A tiny perturbation on a would introduce a dramatic change in x, which makes it too hard to learn the other parameters b, c, d, especially the offset parameter d.

So what are the important points that make the model successfully fit a polyline?

lucastabelini commented 2 years ago

I am not sure that I follow. How would you know the limits of the polyline?