Skip to content


Repository files navigation


This codebase is a TensorFlow implementation of our ECCV-2018 paper:

Recovering 3D Planes from a Single Image via Convolutional Neural Networks

Fengting Yang and Zihan Zhou

Please contact Fengting Yang ( if you have any questions.


This codebase was developed and tested with python 2.7, Tensorflow 1.4.1, CUDA 8.0.61 and Ubuntu 16.04.

Preparing training data

Here we provide our training and testing data on SYNTHIA dataset. Once you download the training data, you can set the training data path as <SYNTHIA_DUMP_DIR> in the training command and start to train the network.

If you wish to generate the training data by yourself, you may want to follow the following steps.

First, download the four-season sequences (Spring, Summer, Fall, Winter) of SEQS-02, SEQS-04, SEQS-05, and save them in one folder <SYNTHIA_DIR>. Then run the following command to filter out the static frames and generate the training data

python data_pre_processing/SYNTHIA/ --dataset_dir=<SYNTHIA_DIR> --dump_root=<SYNTHIA_DUMP_Filtered_DIR> 
python data_pre_processing/SYNTHIA/ --filtered_dataset=<SYNTHIA_DUMP_Filter_DIR> --dump_root=<SYNTHIA_DUMP_DIR> 

The code will generate two ".txt" files for training and testing, we recommend replacing the tst_100.txt with the one in the data_pre_processing/SYNTHIA folder for the availability of the ground truth. The "train_8000.txt" in some folder records the training data we used in our training. Please note the depth unit of SYNTHIA is centimeter, so we divide the depth map by 100.0 in the data loading process.


Once the data is prepared, you should be able to train the model by running the following command

python --dataset_dir=<SYNTHIA_DUMP_DIR> --log_dir=<CKPT_LOG_DIR>

if you want to continue to train or fine-tune from a pre-trained model, you can run

python --dataset_dir=<SYNTHIA_DUMP_DIR> --log_dir=<CKPT_LOG_DIR> --init_checkpoint_file=<PATH_TO_THE_CKPT> --continue_train=True

You can then start a tensorboard session by

tensorboard --logdir=<DIR_CONTAINS_THE_EVENT_FILE> --port=6006

and monitor the training progress by opening the 6006 port on your browser. If everything is set up properly, reasonable segmentation should be observed around 200k steps. The number of recovered planes will keep increasing until it reaches the maximum number set in the code (default=5).

A pre-trained model has been included in the folder named "pre_trained_model", and the ground truth segmentation is in "eval/labels/".


We provide test code to generate: 1) plane segmentation (and its visualization) and 2) depth prediction (planar area only). The evaluation of the depth prediction accuracy will be presented right after the test process. Please run

python --dataset=<SYNTHIA_DUMP_Filtered_DIR or SYNTHIA_DUMP_DIR> --output_dir=<TEST_OUTPUT_DIR> --test_list=<Tst_100.txt in SYNTHIA_DUMP_DIR> --ckpt_file=<TRAINED_MODEL> --use_preprocessed=<True if use our preprocessed dataset, otherwise False>

We also provide code to generate planar 3D models. Please run

cd eval
python --dataset=<SYNTHIA_DUMP_Filtered_DIR or SYNTHIA_DUMP_DIR> --pred_depth=<TEST_OUTPUT_DIR/depth> --savepath=<PATH_TO_SAVE_PLY> --test_list=<Tst_100.txt in SYNTHIA_DUMP_DIR> --use_preprocessed=<True if use our preprocessed dataset, otherwise False>

The point cloud of the .ply file for the planar model should be viewed in <PATH_TO_SAVE_PLY>. You can visualize it through MeshLab directly.


  1. We use the filtered data as input instead of the pre-processed one (to preserve the resolution of the ground truth depth). If you do not want to do the pre-processing and have already downloaded our data, you can simply modify the path related to the dataset in The final result may not be exactly the same as ours but should be similar. And please set --use_preprocessed=True at testing time.
  2. We intentionally exclude seq.22 in our training to test the model performance in a video clip. That is why this sequence is missing in the provided training/test data. We can run the same test by downloading the filtered seq.22 (without pre-processing) here and generate a test list accordingly.


We also provide the MATLAB code for the plane segmentation evaluation:

(1) Open the eval/eval_planes.m;
(2) Set the pred_path as the path to the plane_sgmts folder generated in the test step and check if the label_path is appropriately pointing to the eval/labels/SYN_GT_sgmt;
(3) Run the program, you should be able to see the evaluation result on the command window.


Our code is developed based on the training framework provided by SfMLearner


[ECCV'18] Recovering 3D Planes from a Single Image via Convolutional Neural Networks







No releases published


No packages published