VMLoc: Variational Fusion For Learning-Based Multimodal Camera Localization

Authors

  • Kaichen Zhou University of Oxford
  • Changhao Chen National University of Defense Technology
  • Bing Wang University of Oxford
  • Muhamad Risqi U. Saputra University of Oxford
  • Niki Trigoni University of Oxford
  • Andrew Markham University of Oxford

Keywords:

Localization, Mapping, and Navigation

Abstract

Recent learning-based approaches have achieved impressive results in the field of single-shot camera localization. However, how best to fuse multiple modalities (e.g., image and depth) and to deal with degraded or missing input are less well studied. In particular, we note that previous approaches towards deep fusion do not perform significantly better than models employing a single modality. We conjecture that this is because of the naive approaches to feature space fusion through summation or concatenation which do not take into account the different strengths of each modality. To address this, we propose an end-to-end framework, termed VMLoc, to fuse different sensor inputs into a common latent space through a variational Product-of-Experts (PoE) followed by attention-based fusion. Unlike previous multimodal variational works directly adapting the objective function of vanilla variational auto-encoder, we show how camera localization can be accurately estimated through an unbiased objective function based on importance weighting. Our model is extensively evaluated on RGB-D datasets and the results prove the efficacy of our model. The source code is available at https://github.com/Zalex97/VMLoc.

Downloads

Published

2021-05-18

How to Cite

Zhou, K., Chen, C., Wang, B., Saputra, M. R. U., Trigoni, N., & Markham, A. (2021). VMLoc: Variational Fusion For Learning-Based Multimodal Camera Localization. Proceedings of the AAAI Conference on Artificial Intelligence, 35(7), 6165-6173. Retrieved from https://ojs.aaai.org/index.php/AAAI/article/view/16767

Issue

Section

AAAI Technical Track on Intelligent Robots