JiaRenChang / PSMNet

Pyramid Stereo Matching Network (CVPR2018)
MIT License
1.44k stars 423 forks source link

关于论文中与程序中CNN dilation设置不一致的问题? #182

Open williamma111 opened 4 years ago

williamma111 commented 4 years ago

这是程序中的dilation为 1image 这是论文中dilation image 请问这里应该以哪个为准呢?

JiaRenChang commented 4 years ago

@williamma111

image

We already clarified this setting in the caption in the Table 2. We used half of the dilation.

ltfAutos commented 1 year ago

@williamma111

image

We already clarified this setting in the caption in the Table 2. We used half of the dilation.

May I ask if the weight used in kitti for 1.83 is the 1.09 or 1.12 pre trained weight for scenefrlow? Would the effect be better if the scene flow uses double expansion rate and then the pre trained model kitti is trained with single expansion rate?