Optimizing Through Learned Errors for Accurate Sports Field Registration
arXiv:1909.08034
Abstract
We propose an optimization-based framework to register sports field templates onto broadcast videos. For accurate registration we go beyond the prevalent feed-forward paradigm. Instead, we propose to train a deep network that regresses the registration error, and then register images by finding the registration parameters that minimize the regressed error. We demonstrate the effectiveness of our method by applying it to real-world sports broadcast videos, outperforming the state of the art. We further apply our method on a synthetic toy example and demonstrate that our method brings significant gains even when the problem is simplified and unlimited training data is available.
References in corpus (9)
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
- Spectral Normalization for Generative Adversarial Networks
- TILDE: A Temporally Invariant Learned DEtector
- LF-Net: Learning Local Features from Images
- Deep Image Homography Estimation
- BA-Net: Dense Bundle Adjustment Network
- Deep Value Networks Learn to Evaluate and Iteratively Refine Structured Outputs
- LS-Net: Learning to Solve Nonlinear Least Squares for Monocular Stereo
- RegNet: Learning the Optimization of Direct Image-to-Image Pose Registration