# Estimating 2D Camera Motion with Hybrid Motion Basis

arXiv: https://arxiv.org/abs/2507.22480 · HTML: https://mlsanity.com/paper/2507.22480/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 28 of its 48 references that are in the map.
- First posted on arXiv 30 Jul 2025.

## Builds on (most cited first, 28 of 28)

- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [Segment Anything](https://mlsanity.com/paper/2304.02643/) (2023)
- [SuperGlue: Learning Feature Matching with Graph Neural Networks](https://mlsanity.com/paper/1911.11763/) (2019)
- [SuperPoint: Self-Supervised Interest Point Detection and Description](https://mlsanity.com/paper/1712.07629/) (2017)
- [LoFTR: Detector-Free Local Feature Matching with Transformers](https://mlsanity.com/paper/2104.00680/) (2021)
- [LIFT: Learned Invariant Feature Transform](https://mlsanity.com/paper/1603.09114/) (2016)
- [MAGSAC++, a fast, reliable and accurate robust estimator](https://mlsanity.com/paper/1912.05909/) (2019)
- [GLU-Net: Global-Local Universal Network for Dense Flow and Correspondences](https://mlsanity.com/paper/1912.05524/) (2019)
- [Deep Image Homography Estimation](https://mlsanity.com/paper/1606.03798/) (2016)
- [Uncertainty Estimates and Multi-Hypotheses Networks for Optical Flow](https://mlsanity.com/paper/1802.07095/) (2018)
- [SOSNet: Second Order Similarity Regularization for Local Descriptor Learning](https://mlsanity.com/paper/1904.05019/) (2019)
- [MAGSAC: marginalizing sample consensus](https://mlsanity.com/paper/1803.07469/) (2018)
- [Learning Accurate Dense Correspondences and When to Trust Them](https://mlsanity.com/paper/2101.01710/) (2021)
- [Lightweight Probabilistic Deep Networks](https://mlsanity.com/paper/1805.11327/) (2018)
- [Unsupervised Deep Homography: A Fast and Robust Homography Estimation Model](https://mlsanity.com/paper/1709.03966/) (2017)
- [Content-Aware Unsupervised Deep Homography Estimation](https://mlsanity.com/paper/1909.05983/) (2019)
- [Deep Homography Estimation for Dynamic Scenes](https://mlsanity.com/paper/2004.02132/) (2020)
- [k-Nearest Neighbour Classifiers: 2nd Edition (with Python examples)](https://mlsanity.com/paper/2004.04523/) (2020)
- [RANSAC-Flow: generic two-stage image alignment](https://mlsanity.com/paper/2004.01526/) (2020)
- [Iterative Deep Homography Estimation](https://mlsanity.com/paper/2203.15982/) (2022)
- [Motion Basis Learning for Unsupervised Deep Homography Estimation with Subspace Projection](https://mlsanity.com/paper/2103.15346/) (2021)
- [LocalTrans: A Multiscale Local Transformer Network for Cross-Resolution Homography Estimation](https://mlsanity.com/paper/2106.04067/) (2021)
- [Unsupervised Homography Estimation with Coplanarity-Aware GAN](https://mlsanity.com/paper/2205.03821/) (2022)
- [GyroFlow: Gyroscope-Guided Unsupervised Optical Flow Learning](https://mlsanity.com/paper/2103.13725/) (2021)
- [Semi-supervised Deep Large-baseline Homography Estimation with Progressive Equivalence Constraint](https://mlsanity.com/paper/2212.02763/) (2022)
- [GyroFlow+: Gyroscope-Guided Unsupervised Deep Homography and Optical Flow Learning](https://mlsanity.com/paper/2301.10018/) (2023)
- [DeepMeshFlow: Content Adaptive Mesh Deformation for Robust Image Registration](https://mlsanity.com/paper/1912.05131/) (2019)

## Abstract (arXiv)

Estimating 2D camera motion is a fundamental computer vision task that models the projection of 3D camera movements onto the 2D image plane. Current methods rely on either homography-based approaches, limited to planar scenes, or meshflow techniques that use grid-based local homographies but struggle with complex non-linear transformations. A key insight of our work is that combining flow fields from different homographies creates motion patterns that cannot be represented by any single homography. We introduce CamFlow, a novel framework that represents camera motion using hybrid motion bases: physical bases derived from camera geometry and stochastic bases for complex scenarios. Our approach includes a hybrid probabilistic loss function based on the Laplace distribution that enhances training robustness. For evaluation, we create a new benchmark by masking dynamic objects in existing optical flow datasets to isolate pure camera motion. Experiments show CamFlow outperforms state-of-the-art methods across diverse scenarios, demonstrating superior robustness and generalization in zero-shot settings. Code and datasets are available at our project page: https://lhaippp.github.io/CamFlow/.