Official PyTorch implementation of Focus on Defocus: Bridging the Synthetic to Real Domain Gap for Depth Estimation paper published at Conference on Computer Vision and Pattern Recognition (CVPR) 2020.

Please download the code:

To use our code, first download the repository:

git clone https://github.com/dvl-tum/defocus-net.git

To install the dependencies:

pip install -r requirements.txt

In order to train, run the following command:

python run_training.py

You can find the dataset here.

You would need to upload it to the 'data/' folder. The first 80% of it is the training data (400 samples). Each sample consists of a focal stack with 5 images and a depth file.

Data settings:

Focus distances: [0.1, 0.15, 0.3, 0.7, 1.5]. 
depth_max = 3 (meters)
focal_length = 2.9 * 1e-3
f_number = 1.

If you would like to render your own dataset, we provide a blender file with a python script to render focal stacks with depths (data/scene_focalstack.blend). You would need to prepare object meshes and environment maps (+ optionally textures ).

Objects meshes can be downloaded from Thingi10k dataset: https://ten-thousand-models.appspot.com/

Environment maps can be found at https://hdrihaven.com/

If you find this code useful, please consider citing the following paper:

@InProceedings{Defocus_2020_CVPR,
author = {Maximov, Maxim and Galim, Kevin and Leal-Taixe, Laura},
title = {Focus on Defocus: Bridging the Synthetic to Real Domain Gap for Depth Estimation},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2020}
}