Surface Normals in the Wild

Code for reproducing the results in the following paper:

Surface Normals in the Wild,
Weifeng Chen, Donglai Xiang, Jia Deng
International Conference on Computer Vision (ICCV), 2017.

Please check out the project site for more details.

Setup

1. Install the Torch 7 framework as described in http://torch.ch/docs/getting-started.html#_. Please make sure that you have the cudnn, hdf5, 'mattorch' and csvigo modules installed.

2. Clone this repo.

    https://github.com/wfchen-umich/surface_normals.git
   

Evaluating on pre-trained models

Setup

Please first download the data files and pre-trained models into the surface_normals folder. Download the SNOW dataset from the project site.

Untar data.tar.gz into surface_normals. Untar results.tar.gz into surface_normals/src. Untar SNOW_Toolkit.tar.gz into surface_normals/data. Untar SNOW_images.tar.gz into surface_normals/data/SNOW_Toolkit.

NYU Experiments

Change directory into /surface_normals/src/experiment_NYU.

NYU Subset

To evaluate the pre-trained models ( trained on the NYU labeled training subset), run the following commands:

• d_n_al:

    th test_model_on_NYU_NO_CROP.lua -num_iter 1000 -prev_model_file ../results/hourglass3_softplus_margin_log/wn1_n5000_d800/model_period2_100000.t7 -test_set 654_NYU_MITpaper_test_imgs_orig_size_points.csv -mode test
  

• d_n_dl:

    th test_model_on_NYU_NO_CROP.lua -num_iter 1000 -prev_model_file ../results/hourglass3_softplus_margin_log_depth_from_normal/wn100_n5000_d800/model_period2_100000.t7 -test_set 654_NYU_MITpaper_test_imgs_orig_size_points.csv -mode test
  

NYU Full

To evaluate the pre-trained models ( trained on the full NYU labeled training subset), run the following commands:

• d_n_al_F:

    th test_model_on_NYU_NO_CROP.lua -num_iter 1000 -prev_model_file ../results/hourglass3_softplus_margin_log/wn1_n5000_d10000_fullNYU/model_period3_100000.t7 -test_set 654_NYU_MITpaper_test_imgs_orig_size_points.csv -mode test
  

• d_n_dl_F:

    th test_model_on_NYU_NO_CROP.lua -num_iter 1000 -prev_model_file ../results/hourglass3_softplus_margin_log_depth_from_normal/wn100_n5000_d10000_fullNYU/model_period3_90000.t7 -test_set 654_NYU_MITpaper_test_imgs_orig_size_points.csv -mode test
  

SNOW Experiments

Normals from Predicted Depth:

• d_n_al_F_SNOW

    th test_model_on_SNOW.lua -num_iter 100000 -prev_model_file ../results/hourglass3_softplus_margin_log/SNOW12_from_n5000_d10000_1e-4/model_period3_100000.t7 -mode test
  

KITTI Experiments

Change directory into /surface_normals/src/experiment_KITTI. Run the following commands:

• d:

    th test_model_on_KITTI.lua -num_iter 1000 -prev_model_file ../results/hourglass3_softplus_margin_log_depth_from_normal/KITTI_1e-4_n0_run2_1e-5/model_period10_200000.t7 -test_set eigen_test_files_combine.csv -mode test
  

• d_n_al:

    th test_model_on_KITTI.lua -num_iter 1000 -prev_model_file ../results/hourglass3_softplus_margin_log/KITTI_1e-4_d5000_n5000_run2_1e-5/model_period7_150000.t7 -test_set eigen_test_files_combine.csv -mode test
  

• d_n_dl:

    th test_model_on_KITTI.lua -num_iter 1000 -prev_model_file ../results/hourglass3_softplus_margin_log_depth_from_normal/KITTI_1e-4_n5000_run2_1e-5/model_period7_160000.t7 -test_set eigen_test_files_combine.csv -mode test