Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Connecting the Dots: Learning Representations for Active Monocular Depth Estimation


This repository contains the code for the paper

Connecting the Dots: Learning Representations for Active Monocular Depth Estimation
Gernot Riegler, Yiyi Liao, Simon Donne, Vladlen Koltun, and Andreas Geiger
CVPR 2019

We propose a technique for depth estimation with a monocular structured-light camera, i.e., a calibrated stereo set-up with one camera and one laser projector. Instead of formulating the depth estimation via a correspondence search problem, we show that a simple convolutional architecture is sufficient for high-quality disparity estimates in this setting. As accurate ground-truth is hard to obtain, we train our model in a self-supervised fashion with a combination of photometric and geometric losses. Further, we demonstrate that the projected pattern of the structured light sensor can be reliably separated from the ambient information. This can then be used to improve depth boundaries in a weakly supervised fashion by modeling the joint statistics of image and depth edges. The model trained in this fashion compares favorably to the state-of-the-art on challenging synthetic and real-world datasets. In addition, we contribute a novel simulator, which allows to benchmark active depth prediction algorithms in controlled conditions.

If you find this code useful for your research, please cite

  title={Connecting the Dots: Learning Representations for Active Monocular Depth Estimation},
  author={Riegler, Gernot and Liao, Yiyi and Donne, Simon and Koltun, Vladlen and Geiger, Andreas},
  booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},


The network training/evaluation code is based on Pytorch.


Updated on 07.06.2021: The code is now compatible with the latest Pytorch version (1.8).

The other python packages can be installed with anaconda:

conda install --file requirements.txt

Structured Light Renderer

To train and evaluate our method in a controlled setting, we implemented an structured light renderer. It can be used to render a virtual scene (arbitrary triangle mesh) with the structured light pattern projected from a customizable projector location. To build it, first make sure the correct CUDA_LIBRARY_PATH is set in config.json. Afterwards, the renderer can be build by running make within the renderer directory.

PyTorch Extensions

The network training/evaluation code is based on PyTorch. We implemented some custom layers that need to be built in the torchext directory. Simply change into this directory and run

python build_ext --inplace

Baseline HyperDepth

As baseline we partially re-implemented the random forest based method HyperDepth. The code resided in the hyperdepth directory and is implemented in C++11 with a Python wrapper written in Cython. To build it change into the directory and run

python build_ext --inplace


Creating Synthetic Data

To create synthetic data and save it locally, download ShapeNet V2 and correct SHAPENET_ROOT in config.json. Then the data can be generated and saved to DATA_ROOT in config.json by running


If you are only interested in evaluating our pre-trained model, here (3.7G) is a validation set that contains a small amount of images.

Training Network

As a first stage, it is recommended to train the disparity decoder and edge decoder without the geometric loss. To train the network on synthetic data for the first stage run


After the model is pretrained without the geometric loss, the full model can be trained from the initialized weights by running

python --loss phge

Evaluating Network

To evaluate a specific checkpoint, e.g. the 50th epoch, one can run

python --cmd retest --epoch 50

Evaluating a Pre-trained Model

We provide a model pre-trained using the photometric loss. Once you have prepared the synthetic dataset and changed DATA_ROOT in config.json, the pre-trained model can be evaluated on the validation set by running:

mkdir -p output
mkdir -p output/exp_syn
wget -O output/exp_syn/net_0099.params
python --cmd retest --epoch 99

You can also download our validation set from here (3.7G).


This work was supported by the Intel Network on Intelligent Systems.


This repository contains the code for the paper "Connecting the Dots: Learning Representations for Active Monocular Depth Estimation"







No releases published


No packages published