Salvato in:
Dettagli Bibliografici
Autori principali: Yang, Jinghe, Gong, Mingming, Nair, Girish, Lee, Jung Hoon, Monty, Jason, Pu, Ye
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:https://arxiv.org/abs/2303.17981
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909090447360000
author Yang, Jinghe
Gong, Mingming
Nair, Girish
Lee, Jung Hoon
Monty, Jason
Pu, Ye
author_facet Yang, Jinghe
Gong, Mingming
Nair, Girish
Lee, Jung Hoon
Monty, Jason
Pu, Ye
contents In recent years, learning-based feature detection and matching have outperformed manually-designed methods in in-air cases. However, it is challenging to learn the features in the underwater scenario due to the absence of annotated underwater datasets. This paper proposes a cross-modal knowledge distillation framework for training an underwater feature detection and matching network (UFEN). In particular, we use in-air RGBD data to generate synthetic underwater images based on a physical underwater imaging formation model and employ these as the medium to distil knowledge from a teacher model SuperPoint pretrained on in-air images. We embed UFEN into the ORB-SLAM3 framework to replace the ORB feature by introducing an additional binarization layer. To test the effectiveness of our method, we built a new underwater dataset with groundtruth measurements named EASI (https://github.com/Jinghe-mel/UFEN-SLAM), recorded in an indoor water tank for different turbidity levels. The experimental results on the existing dataset and our new dataset demonstrate the effectiveness of our method.
format Preprint
id arxiv_https___arxiv_org_abs_2303_17981
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Knowledge Distillation for Feature Extraction in Underwater VSLAM
Yang, Jinghe
Gong, Mingming
Nair, Girish
Lee, Jung Hoon
Monty, Jason
Pu, Ye
Computer Vision and Pattern Recognition
In recent years, learning-based feature detection and matching have outperformed manually-designed methods in in-air cases. However, it is challenging to learn the features in the underwater scenario due to the absence of annotated underwater datasets. This paper proposes a cross-modal knowledge distillation framework for training an underwater feature detection and matching network (UFEN). In particular, we use in-air RGBD data to generate synthetic underwater images based on a physical underwater imaging formation model and employ these as the medium to distil knowledge from a teacher model SuperPoint pretrained on in-air images. We embed UFEN into the ORB-SLAM3 framework to replace the ORB feature by introducing an additional binarization layer. To test the effectiveness of our method, we built a new underwater dataset with groundtruth measurements named EASI (https://github.com/Jinghe-mel/UFEN-SLAM), recorded in an indoor water tank for different turbidity levels. The experimental results on the existing dataset and our new dataset demonstrate the effectiveness of our method.
title Knowledge Distillation for Feature Extraction in Underwater VSLAM
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2303.17981