Understanding the Limitations of CNN-based Absolute Camera Pose Regression
arXiv:1903.07504
Abstract
Visual localization is the task of accurate camera pose estimation in a known scene. It is a key problem in computer vision and robotics, with applications including self-driving cars, Structure-from-Motion, SLAM, and Mixed Reality. Traditionally, the localization problem has been tackled using 3D geometry. Recently, end-to-end approaches based on convolutional neural networks have become popular. These methods learn to directly regress the camera pose from an input image. However, they do not achieve the same level of pose accuracy as 3D structure-based methods. To understand this behavior, we develop a theoretical model for camera pose regression. We use our model to predict failure cases for pose regression techniques and verify our predictions through experiments. We furthermore use our model to show that pose regression is more closely related to pose approximation via image retrieval than to accurate pose estimation via 3D structure. A key result is that current approaches do not consistently outperform a handcrafted image retrieval baseline. This clearly shows that additional research is needed before pose regression algorithms are ready to compete with structure-based methods.
Initial version of a paper accepted to CVPR 2019
Cited by in corpus (5)
- Introduction to Camera Pose Estimation with Deep Learning
- Lending Orientation to Neural Networks for Cross-view Geo-localization
- Ego-Motion Alignment from Face Detections for Collaborative Augmented Reality
- Unconstrained Matching of 2D and 3D Descriptors for 6-DOF Pose Estimation
- Deep Weakly Supervised Positioning