kbatsos / CBMV

Welcome to the code repository of CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. CVPR 2018
GNU General Public License v3.0
53 stars 13 forks source link

question #7

Open yangguanxun opened 5 years ago

yangguanxun commented 5 years ago

配置好环境之后,如何实现这个项目?????? 望大神不吝赐教

kbatsos commented 5 years ago

Please follow the instruction on README

yangguanxun commented 5 years ago

scipy的版本也有要求吗

kbatsos commented 5 years ago

sklearn (version 0.17.1). Please read the README. scipy recent should work

ccj5351 commented 5 years ago

配置好环境之后,如何实现这个项目?????? 望大神不吝赐教

Yes. Just follow the Compilation section to compile the code, and then Training section for random forest training and Disparity Estimation section for disparity map as output.

ccj5351 commented 5 years ago

We specify the sklearn as version 0.17.1, it is because that the model we provided here is trained in this setting. If you use newer version sklearn to load our model, there might exist some problems. But you can always use new version sklearn as long as the training and testing both using this version.

yangguanxun commented 5 years ago

这种问题该如何解决 Traceback (most recent call last): File "main.py", line 19, in from train import Training as cbmvtrain File "./train.py", line 25, in import librectification as rect ImportError: No module named librectification

kbatsos commented 5 years ago

Go to the cpp folder and compiler the relevant libraries. The instructions are on the README please follow the instructions.

kbatsos commented 5 years ago

If you don't want to use a GPU, don't compile post and also remove the post processing pipeline from the python code.