Giving lane prediction even if a blank image is passed.

lucastabelini / PolyLaneNet

Code for the paper entitled "PolyLaneNet: Lane Estimation via Deep Polynomial Regression" (ICPR 2020)

https://arxiv.org/abs/2004.10924

MIT License

301 stars 75 forks source link

Giving lane prediction even if a blank image is passed. #18

Closed Maheshtaparia01 closed 4 years ago

Maheshtaparia01 commented 4 years ago

Hi The model is predicting the lane even if the nothing is there (image with all pixel with 0 / 1 value) or if I give a cat image as input, I am getting 3-4 lanes on that.

lucastabelini commented 4 years ago

I see two reasons for that. One is that these kinds of images are very different from the ones in the training set, which means the output is nothing meaningful. The other is that all datasets used on the paper have at least one lane per image. If the model you used was trained on TuSimple, it will have a huge bias towards predicting 3+ lanes, since the vast majority of the images have this number of lanes.