FlowNorm: A Learning-based Method for Increasing Convergence Range of Direct Alignment

TL;DR

FlowNorm is a novel learning-based robust norm that combines local patch alignment and global image registration information to significantly enhance the convergence range of direct alignment methods for camera pose estimation.

Contribution

The paper introduces FlowNorm, a new robust norm leveraging both local and global information, improving convergence range in direct alignment for camera pose estimation.

Findings

Achieves unprecedented convergence range with large view angle changes.

Integrates into DSO and BA-Net for more robust, real-time results.

Enhances robustness and accuracy of direct alignment methods.

Abstract

Many approaches have been proposed to estimate camera poses by directly minimizing photometric error. However, due to the non-convex property of direct alignment, proper initialization is still required for these methods. Many robust norms (e.g. Huber norm) have been proposed to deal with the outlier terms caused by incorrect initializations. These robust norms are solely defined on the magnitude of each error term. In this paper, we propose a novel robust norm, named FlowNorm, that exploits the information from both the local error term and the global image registration information. While the local information is defined on patch alignments, the global information is estimated using a learning-based network. Using both the local and global information, we achieve an unprecedented convergence range in which images can be aligned given large view angle changes or small overlaps. We further demonstrate the usability of the proposed robust norm by integrating it into the direct methods DSO and BA-Net, and generate more robust and accurate results in real-time.