# Conditional Prior Networks for Optical Flow

arXiv: https://arxiv.org/abs/1807.10378 · HTML: https://mlsanity.com/paper/1807.10378/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 26 Jul 2018.

## Built on by (most cited first, 11 of 11)

- [Reliable amortized variational inference with physics-based latent distribution correction](https://mlsanity.com/paper/2207.11640/) (2022)
- [Preconditioned training of normalizing flows for variational inference in inverse problems](https://mlsanity.com/paper/2101.03709/) (2021)
- [VOLDOR: Visual Odometry from Log-logistic Dense Optical flow Residuals](https://mlsanity.com/paper/2104.06789/) (2021)
- [Stereoscopic Universal Perturbations across Different Architectures and Datasets](https://mlsanity.com/paper/2112.06116/) (2021)
- [Optical Flow Training under Limited Label Budget via Active Learning](https://mlsanity.com/paper/2203.05053/) (2022)
- [Self-supervised AutoFlow](https://mlsanity.com/paper/2212.01762/) (2022)
- [Tumor aware recurrent inter-patient deformable image registration of computed tomography scans with lung cancer](https://mlsanity.com/paper/2409.11910/) (2024)
- [Divided Attention: Unsupervised Multi-Object Discovery with Contextually Separated Slots](https://mlsanity.com/paper/2304.01430/) (2023)
- [Triangular Consistency as a Universal Constraint for Learning Optical Flow](https://mlsanity.com/paper/2606.19938/) (2026)
- [Scene Uncertainty and the Wellington Posterior of Deterministic Image Classifiers](https://mlsanity.com/paper/2106.13870/) (2021)
- [On the Viability of Monocular Depth Pre-training for Semantic Segmentation](https://mlsanity.com/paper/2203.13987/) (2022)

## Abstract (arXiv)

Classical computation of optical flow involves generic priors (regularizers) that capture rudimentary statistics of images, but not long-range correlations or semantics. On the other hand, fully supervised methods learn the regularity in the annotated data, without explicit regularization and with the risk of overfitting. We seek to learn richer priors on the set of possible flows that are statistically compatible with an image. Once the prior is learned in a supervised fashion, one can easily learn the full map to infer optical flow directly from two or more images, without any need for (additional) supervision. We introduce a novel architecture, called Conditional Prior Network (CPN), and show how to train it to yield a conditional prior. When used in conjunction with a simple optical flow architecture, the CPN beats all variational methods and all unsupervised learning-based ones using the same data term. It performs comparably to fully supervised ones, that however are fine-tuned to a particular dataset. Our method, on the other hand, performs well even when transferred between datasets.