PatchmatchNet

PatchmatchNet copied to clipboard

训练DTU数据的loss问题

24

感谢作者的开源代码！我在尝试训练DTU数据时（新版训练和legacy训练），总是在patchmatchnet_loss函数中报错，报错信息如下： File "/media/user/Data/patchmatchnet_improve/PatchmatchNet-main/models/net.py", line 341, in patchmatchnet_loss loss = loss + F.smooth_l1_loss(depth[mask[i]], gt_depth, reduction="mean") IndexError: The shape of the mask [2, 1, 84, 80] at index 2 does not match...

Be997398715

invalid input folder

1

hi there,thanks for your excellent work . I 'm a nerd in .net work,I download the DTU ,but when i try the sh eval.sh ,the problems came out like invalid...

gaofang-xsy

GPU memory consumption

17

Hi, when running PatchmatchNet on ETH3D dataset through `eval.py`, I end up using 15GB of GPU memory, and the paper reports 5529 MB. Could it be the case that all...

vevenom

train on ETH3D dataset

11

Hi, I downloaded the ETH3D dataset from MVSNET repo and use 'convet_eth3d_dataset.py' to process it. However, i encountered an error during the training process. Traceback (most recent call last): File...

TWang1017

Why is it that when I tested the eth3d dataset using your preprocessed dataset, the obtained completeness, precision, and F1 scores are all very low?

cangmis

I tried it on my own dataset, but the depth maps look bad

3

Thank you very much for providing this code. I was able to get it running pretty smoothly with the instructions provided. The fused.ply look fine, but the depth maps look...

jlartois

Training existing data sets

1

It seems that few existing datasets can be used to train patchmatchnet. Even those datasets used for evaluation (such as ETH3D) do not provide enough information. I wonder if anyone...

LIN-FZ

Hi! Thanks for sharing your work. I'm trying to integrate PatchmatchNet with colmap, at stage stereo_fusion. My input to colmap is dense folder generated with colmap_output.py from dtu/scan1 ``` ....

q-eldar

训练DTU的eval点云问题

17

我使用你提供的ckpt文件输出的点云结果不佳，所以使用train_dtu重新训练了新的参数，loss converge to 7~46，但是输出的点云效果不佳，测试数据集为DTU，请问evaluation有什么要求吗

HDjpf

I use the parameters you provided to run script,but do not get meaningful result. Then I train model using DTU's training set.But new parameters are not meaningful. testing data is...

HDjpf