Skip to content

Code for the paper "MICL: Mutual Information Guided Continual Learning for LiDAR Place Recognition"

Notifications You must be signed in to change notification settings

npu-ius-lab/MICL

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 

Repository files navigation

MICL

Code for the paper "MICL: Mutual Information Guided Continual Learning for LiDAR Place Recognition".

Pretrained Models

The following models from the paper are provided for evaluation purposes, all the models are firstly trained on the oxford dataset and continually trained on the existing three ones. The results of different backbone networks using Mean recall@1 metric are as follows:

Metric learning-based MICL

Architecture Mean Recall@1
MinkLoc3D 85.9
LoGG3D-Net 80.3
PointNetVLAD 65.6

Contrastive learning-based MICL

Architecture Mean Recall@1
MinkLoc3D 84.1
LoGG3D-Net 85.2
PointNetVLAD 68.2

Checkpoints Link: https://drive.google.com/drive/folders/1l2v7Jw3PpIQmN5jjrZntPwzHwTVT5JD7?usp=drive_link

About

Code for the paper "MICL: Mutual Information Guided Continual Learning for LiDAR Place Recognition"

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published