Self-Supervised Learning for Stereo Reconstruction on Aerial Images

Patrick Knöbelreiter, Christoph Vogel, Thomas Pock

Research output: Contribution to conferencePaperpeer-review


Recent developments established deep learning as an inevitable tool to boost the performance of dense matching and stereo estimation. On the downside, learning these networks requires a substantial amount of training data to be successful. Consequently, the application of these models outside of the laboratory is far from straight forward. In this work we propose a self-supervised training procedure that allows us
to adapt our network to the specific (imaging) characteristics of the dataset at hand, without the requirement of external ground truth data. We instead generate interim training data by running our intermediate network on the whole dataset, followed by conservative outlier filtering. Bootstrapped from a pre-trained version of our hybrid CNN-CRF model, we alternate the generation of training data and network training.
With this simple concept we are able to lift the completeness and accuracy of the pretrained version significantly. We also show that our final model compares favorably to other popular stereo estimation algorithms on an aerial dataset.
Original languageEnglish
Publication statusPublished - 22 Jul 2018
Event38th Annual IEEE International Geoscience and Remote Sensing Symposium: IGARSS 2018 - Valencia, Valencia, Spain
Duration: 22 Jul 201827 Jul 2018


Conference38th Annual IEEE International Geoscience and Remote Sensing Symposium
Abbreviated titleIGARSS
Internet address


  • large scale 3D
  • dense matching
  • CNN


Dive into the research topics of 'Self-Supervised Learning for Stereo Reconstruction on Aerial Images'. Together they form a unique fingerprint.

Cite this