xjqi / geonet Goto Github PK

Hi,
Thanks for sharing GeoNet code here! Can you kindly offer NYUD train+test surface normal annotations mentioned in your paper with a direct link (e.g. google drive) ? Compared to the existing ones, they looks better and can serve as a new source of NYUD surface normal estimation training data. Thank you by advance!

The compressed file is too big, the download process is often broken off.
Could you please divide the data into several small files and let us download?
Thanks a lot!

Hi thanks for great work,I wonder will you plan to release the code of train in the near future?

best wishes

hi, thanks for the great work!

when downloading the pretrained network from google drive and the data from hkuhk-my.sharepoint.com, I failed on my Linux server cause both the data were too large and there were VPN issues as I am in china.

hope you can give a seminated data package that divides the data into several small files and help in downloading, if possible.

thanks again!

Hi,
I was trying to train your GeoNet model and I find the training set is from the address in traindata_grid.txt. But in this file I found a list of local address like this:

/home/xjqi/xjqi_storage/depthnorm/depth_norm_cvpr18/data_grid/basement_0001a_r-1316653580.484909-1316500621.mat
/home/xjqi/xjqi_storage/depthnorm/depth_norm_cvpr18/data_grid/basement_0001a_r-1316653580.838319-1338503784.mat

Would you tell me where can I download this dataset please? Thank you very much.

I noticed that you read grid from data_dic['grid'], and data_dic is loaded from files which are included in ./list/traindata_grid.txt. But you didn't provide the training data.
So, what does "grid" mean in your code?

Thanks.

Hey @xjqi , thanks for sharing the code. When I ran the code.py, I have updated both the prediction files (depth_pred.mat and norm_pred.mat) and estimation files(depth_estimate.mat and norm_estimate.mat) for the 654 test images provided, to see the comparison in the evaluation between the predicted output (initial depth and normal) and estimated output (final depth and normal). However, after running the evaluation, I have found that the accuracy for prediction (around 98.9%) is more than the estimation (around 98.3%). I ran the test for the test_images for another 4 iteration and, even though the accuracy for prediction remains the same, the estimation is still lesser (around 98.3%). When i tried to visualise the depth and normal maps, the estimation seems to be more distorted than the depth prediction. Please suggest a solution for this problem or any step I have done wrong.

Hi!Thanks for your code.Would you please share the pytorch version about this project?