Continuous Gesture Sequences Recognition Based on Few-Shot Learning

A large number of demands for space on-orbit services to ensure the on-orbit system completes its specified tasks are foreseeable, and the efficiency and the security are the most significant factors when we carry out an on-orbit mission. And it can improve human-computer interaction efficiency in o...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhe Liu, Cao Pan, Hongyuan Wang
Format: Article
Language:English
Published: Wiley 2022-01-01
Series:International Journal of Aerospace Engineering
Online Access:http://dx.doi.org/10.1155/2022/7868142
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850173863750533120
author Zhe Liu
Cao Pan
Hongyuan Wang
author_facet Zhe Liu
Cao Pan
Hongyuan Wang
author_sort Zhe Liu
collection DOAJ
description A large number of demands for space on-orbit services to ensure the on-orbit system completes its specified tasks are foreseeable, and the efficiency and the security are the most significant factors when we carry out an on-orbit mission. And it can improve human-computer interaction efficiency in operations with proper gesture recognition solutions. In actual situations, the operations are complex and changeable, so the gestures used in interaction are also difficult to predict in advance due to the compounding of multiple consecutive gestures. To recognize such gestures based on computer vision (CV) requires complex models trained by a large amount of datasets, it is often unable to obtain enough gesture samples for training a complex model in real tasks, and the cost of labeling the collected gesture samples is quite expensive. Aiming at the problems mentioned above, we propose a few-shot continuous gesture recognition scheme based on RGB video. The scheme uses Mediapipe to detect the key points of each frame in the video stream, decomposes the basic components of gesture features based on certain human palm structure, and then extracts and combines the above basic gesture features by a lightweight autoencoder network. Our scheme can achieve 89.73% recognition accuracy on the 5-way 1-shot gesture recognition task which randomly selected 142 gesture instances of 5 categories from the RWTH German fingerspelling dataset.
format Article
id doaj-art-d40fefeb8e1a4dc9bdc91aa16e8918a5
institution OA Journals
issn 1687-5974
language English
publishDate 2022-01-01
publisher Wiley
record_format Article
series International Journal of Aerospace Engineering
spelling doaj-art-d40fefeb8e1a4dc9bdc91aa16e8918a52025-08-20T02:19:46ZengWileyInternational Journal of Aerospace Engineering1687-59742022-01-01202210.1155/2022/7868142Continuous Gesture Sequences Recognition Based on Few-Shot LearningZhe Liu0Cao Pan1Hongyuan Wang2Changzhou UniversityChangzhou UniversityChangzhou UniversityA large number of demands for space on-orbit services to ensure the on-orbit system completes its specified tasks are foreseeable, and the efficiency and the security are the most significant factors when we carry out an on-orbit mission. And it can improve human-computer interaction efficiency in operations with proper gesture recognition solutions. In actual situations, the operations are complex and changeable, so the gestures used in interaction are also difficult to predict in advance due to the compounding of multiple consecutive gestures. To recognize such gestures based on computer vision (CV) requires complex models trained by a large amount of datasets, it is often unable to obtain enough gesture samples for training a complex model in real tasks, and the cost of labeling the collected gesture samples is quite expensive. Aiming at the problems mentioned above, we propose a few-shot continuous gesture recognition scheme based on RGB video. The scheme uses Mediapipe to detect the key points of each frame in the video stream, decomposes the basic components of gesture features based on certain human palm structure, and then extracts and combines the above basic gesture features by a lightweight autoencoder network. Our scheme can achieve 89.73% recognition accuracy on the 5-way 1-shot gesture recognition task which randomly selected 142 gesture instances of 5 categories from the RWTH German fingerspelling dataset.http://dx.doi.org/10.1155/2022/7868142
spellingShingle Zhe Liu
Cao Pan
Hongyuan Wang
Continuous Gesture Sequences Recognition Based on Few-Shot Learning
International Journal of Aerospace Engineering
title Continuous Gesture Sequences Recognition Based on Few-Shot Learning
title_full Continuous Gesture Sequences Recognition Based on Few-Shot Learning
title_fullStr Continuous Gesture Sequences Recognition Based on Few-Shot Learning
title_full_unstemmed Continuous Gesture Sequences Recognition Based on Few-Shot Learning
title_short Continuous Gesture Sequences Recognition Based on Few-Shot Learning
title_sort continuous gesture sequences recognition based on few shot learning
url http://dx.doi.org/10.1155/2022/7868142
work_keys_str_mv AT zheliu continuousgesturesequencesrecognitionbasedonfewshotlearning
AT caopan continuousgesturesequencesrecognitionbasedonfewshotlearning
AT hongyuanwang continuousgesturesequencesrecognitionbasedonfewshotlearning