This is an old revision of the document!
GN-Net: The Gauss-Newton Loss for Multi-Weather Relocalization
Contact: Lukas von Stumberg, News, Qadeer Khan, Prof. Daniel Cremers
Abstract
Direct SLAM methods have shown exceptional performance on odometry tasks. However, they are susceptible to dynamic lighting and weather changes while also suffering from a bad initialization on large baselines. To overcome this, we propose GN-Net: a network optimized with the novel Gauss-Newton loss for training weather invariant deep features, tailored for direct image alignment. Our network can be trained with pixel correspondences between images even from different sequences. Experiments on both simulated and real-world datasets demonstrate that our approach is more robust against bad initialization, variations in day-time, and weather changes thereby outperforming state-of-the-art direct and indirect methods. Furthermore, we release an evaluation benchmark for relocalization tracking against different types of weather.
Downloads
The paper can be downloaded at: https://arxiv.org/abs/1904.11932
The video is available at: https://youtu.be/gcbKeKX2eiE
The supplementary can be downloaded at: gn-net-supplementary.pdf
The relocalization tracking benchmark dataset can be downloaded at:
gnnet_benchmark_v1.0.zip
New Results Oxford Robotcar
The following new results include comparisons to D2-Net and SuperPoint. These keypoint-based methods were designed to be used in combination with the PnP algorithm in a RANSAC scheme. We also show results for a GN-Net model which was only trained on the synthetic CARLA benchmark, but tested on the Oxford sequences (dashed green).
Results Sunny-Overcast
Results Sunny-Rainy
Results Sunny-Snowy
Results Overcast-Rainy
Results Overcast-Snowy
Results Rainy-Snowy
Export as PDF, XML, TEX or BIB
Journal Articles
2020
[] GN-Net: The Gauss-Newton Loss for Multi-Weather Relocalization , In IEEE Robotics and Automation Letters (RA-L), volume 5, 2020. ([arXiv][video][project page][supplementary])