search

FunkyKoki / Look_At_Boundary_PyTorch

A PyTorch re-implement of CVPR 2018 LAB(Look At Bounday) --Champagne Jin
9 stars 1 forks source link

about performance #2

Closed 738654805 closed 5 years ago

738654805 commented 5 years ago

The performance on the 300W is not as good as the results in the paper. The best result i got is about 5.1NME (inter-pupil) on the 300W. Do you have some suggestions? Thx in advance.

guoguoguilai commented 5 years ago

The performance on the 300W is not as good as the results in the paper. The best result i got is about 5.1NME (inter-pupil) on the 300W. Do you have some suggestions? Thx in advance. Did you use the code that @FunkyKoki provided?

738654805 commented 5 years ago

@guoguoguilai I just use the boundary heatmaps re-implement by myself, it seems not work on the 300W.

FunkyKoki commented 5 years ago

Not very well on 300W as well, but better on WFLW(at least close to the result the author gave), later, I would try AFLW and COFW. The performance would be reported in about one month. Thanks for your concern...