Multiple Rotation Averaging with Constrained Reweighting Deep Matrix Factorization

TL;DR

This paper introduces a novel deep matrix factorization approach for multiple rotation averaging that is robust to outliers and does not require ground truth labels, combining optimization and learning techniques.

Contribution

It proposes a label-free, deep matrix factorization method with reweighting and dynamic depth strategies for robust rotation averaging in computer vision.

Findings

Effective outlier suppression via spanning tree filtering

Robust performance demonstrated on multiple datasets

Combines advantages of optimization and learning methods

Abstract

Multiple rotation averaging plays a crucial role in computer vision and robotics domains. The conventional optimization-based methods optimize a nonlinear cost function based on certain noise assumptions, while most previous learning-based methods require ground truth labels in the supervised training process. Recognizing the handcrafted noise assumption may not be reasonable in all real-world scenarios, this paper proposes an effective rotation averaging method for mining data patterns in a learning manner while avoiding the requirement of labels. Specifically, we apply deep matrix factorization to directly solve the multiple rotation averaging problem in unconstrained linear space. For deep matrix factorization, we design a neural network model, which is explicitly low-rank and symmetric to better suit the background of multiple rotation averaging. Meanwhile, we utilize a spanning tree-based edge filtering to suppress the influence of rotation outliers. What's more, we also adopt a reweighting scheme and dynamic depth selection strategy to further improve the robustness. Our method synthesizes the merit of both optimization-based and learning-based methods. Experimental results on various datasets validate the effectiveness of our proposed method.