Adversarial Transfer of Pose Estimation Regression

TL;DR

This paper proposes a novel adversarial transfer learning approach for camera pose estimation that enhances model generalization across different scenes by learning scene-invariant representations.

Contribution

It extends domain adaptation techniques to multi-regression pose estimation using adversarial learning and self-supervised methods for improved cross-scene transfer.

Findings

Outperforms baseline methods on Cambridge Landmarks and 7Scene datasets.

Demonstrates improved generalization across different scenes.

Validates scene-invariant representations using adaptability theory.

Abstract

We address the problem of camera pose estimation in visual localization. Current regression-based methods for pose estimation are trained and evaluated scene-wise. They depend on the coordinate frame of the training dataset and show a low generalization across scenes and datasets. We identify the dataset shift an important barrier to generalization and consider transfer learning as an alternative way towards a better reuse of pose estimation models. We revise domain adaptation techniques for classification and extend them to camera pose estimation, which is a multi-regression task. We develop a deep adaptation network for learning scene-invariant image representations and use adversarial learning to generate such representations for model transfer. We enrich the network with self-supervised learning and use the adaptability theory to validate the existence of scene-invariant representation of images in two given scenes. We evaluate our network on two public datasets, Cambridge Landmarks and 7Scene, demonstrate its superiority over several baselines and compare to the state of the art methods.