Affinity Relation-aware Fine-grained Bird Image Recognition for Robot Vision Tracking via Transformers

Hai Liu; Cheng Zhang; Bochen Xie; Tingting Liu; Qingsong Xu; You-fu Li

doi:10.1109/ROBIO55434.2022.10011861

Affinity Relation-aware Fine-grained Bird Image Recognition for Robot Vision Tracking via Transformers

Hai Liu, Cheng Zhang^*, Bochen Xie, Tingting Liu, Qingsong Xu, You-fu Li^*

^*Corresponding author for this work

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

5 Citations (Scopus)

Abstract

Fine-grained Bird Image Recognition dedicates to achieve accurate bird image classification and it is also a fundamental task in robot vision tracking. Given that endangered bird surveillance and conservation is of great significance for bird protection from extinction, automated approaches are in need to facilitate bird surveillance. In this work, we propose a novel robot vision tracking based method for bird surveillance with an affinity relation-aware model named TBNet that combines CNN and Transformer architecture and has a novel feature selection (FS) module. Specifically, CNN is employed to extract superficial information. Transformer is utilized for exploiting abstract semantic affinity relations. FS module is introduced to reveal discriminative features. Comprehensive experiments demonstrate that can achieve state-of-the-art performance on the CUB-200-2011 dataset (91.0%) and the NABirds dataset (90.9%). © 2022 IEEE

Original language	English
Title of host publication	2022 IEEE International Conference on Robotics and Biomimetics (ROBIO)
Publisher	IEEE
Pages	662-667
ISBN (Electronic)	978-1-6654-8109-0
ISBN (Print)	978-1-6654-8110-6
DOIs	https://doi.org/10.1109/ROBIO55434.2022.10011861
Publication status	Published - Dec 2022
Event	2022 IEEE International Conference on Robotics and Biomimetics (ROBIO 2022) - China Sheraton Grand Xishuangbanna Hotel, Jinghong, China Duration: 5 Dec 2022 → 9 Dec 2022 http://irmv.sjtu.edu.cn/robio2022/venue.html

Conference

Conference	2022 IEEE International Conference on Robotics and Biomimetics (ROBIO 2022)
Abbreviated title	IEEE ROBIO 2022
Country/Territory	China
City	Jinghong
Period	5/12/22 → 9/12/22
Internet address	http://irmv.sjtu.edu.cn/robio2022/venue.html

Funding

This work was supported in part by the National Natural Science Foundation of China under Grant 62277041, Grant 62211530433 Grant 62177018, Grant 62011530436, Grant 62077020, Grant 62005092, and Grant 61875068, the Research Grants Council of Hong Kong under Project 9043323, CityU 11213420, and the Science and Technology Development Fund, Macau, under Grant 0022/2019/AKP

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access Output

10.1109/ROBIO55434.2022.10011861

https://ieeexplore.ieee.org/document/10011861/

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

1 Finished
1 Active

GRF: Event-Based 3D Human Pose Estimation and Tracking
LI, Y. F. (Principal Investigator / Project Coordinator)
1/01/23 → …
Project: Research
GRF: Gaze Tracking and its Integration with Human-Robot Cooperation
LI, Y. F. (Principal Investigator / Project Coordinator) & CHEN, H. (Co-Investigator)
1/01/21 → 24/06/25
Project: Research

Cite this

@inproceedings{a7913d51ed004296b1aa8cab85e1cddb,

title = "Affinity Relation-aware Fine-grained Bird Image Recognition for Robot Vision Tracking via Transformers",

abstract = "Fine-grained Bird Image Recognition dedicates to achieve accurate bird image classification and it is also a fundamental task in robot vision tracking. Given that endangered bird surveillance and conservation is of great significance for bird protection from extinction, automated approaches are in need to facilitate bird surveillance. In this work, we propose a novel robot vision tracking based method for bird surveillance with an affinity relation-aware model named TBNet that combines CNN and Transformer architecture and has a novel feature selection (FS) module. Specifically, CNN is employed to extract superficial information. Transformer is utilized for exploiting abstract semantic affinity relations. FS module is introduced to reveal discriminative features. Comprehensive experiments demonstrate that can achieve state-of-the-art performance on the CUB-200-2011 dataset (91.0%) and the NABirds dataset (90.9%). {\textcopyright} 2022 IEEE",

author = "Hai Liu and Cheng Zhang and Bochen Xie and Tingting Liu and Qingsong Xu and You-fu Li",

year = "2022",

month = dec,

doi = "10.1109/ROBIO55434.2022.10011861",

language = "English",

isbn = "978-1-6654-8110-6",

pages = "662--667",

booktitle = "2022 IEEE International Conference on Robotics and Biomimetics (ROBIO)",

publisher = "IEEE",

address = "United States",

note = "2022 IEEE International Conference on Robotics and Biomimetics (ROBIO 2022), IEEE ROBIO 2022 ; Conference date: 05-12-2022 Through 09-12-2022",

url = "http://irmv.sjtu.edu.cn/robio2022/venue.html",

}

Liu, H, Zhang, C, Xie, B, Liu, T, Xu, Q & Li, Y 2022, Affinity Relation-aware Fine-grained Bird Image Recognition for Robot Vision Tracking via Transformers. in 2022 IEEE International Conference on Robotics and Biomimetics (ROBIO). IEEE, pp. 662-667, 2022 IEEE International Conference on Robotics and Biomimetics (ROBIO 2022), Jinghong, China, 5/12/22. https://doi.org/10.1109/ROBIO55434.2022.10011861

Affinity Relation-aware Fine-grained Bird Image Recognition for Robot Vision Tracking via Transformers. / Liu, Hai; Zhang, Cheng; Xie, Bochen et al.
2022 IEEE International Conference on Robotics and Biomimetics (ROBIO). IEEE, 2022. p. 662-667.

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - Affinity Relation-aware Fine-grained Bird Image Recognition for Robot Vision Tracking via Transformers

AU - Liu, Hai

AU - Zhang, Cheng

AU - Xie, Bochen

AU - Liu, Tingting

AU - Xu, Qingsong

AU - Li, You-fu

PY - 2022/12

Y1 - 2022/12

N2 - Fine-grained Bird Image Recognition dedicates to achieve accurate bird image classification and it is also a fundamental task in robot vision tracking. Given that endangered bird surveillance and conservation is of great significance for bird protection from extinction, automated approaches are in need to facilitate bird surveillance. In this work, we propose a novel robot vision tracking based method for bird surveillance with an affinity relation-aware model named TBNet that combines CNN and Transformer architecture and has a novel feature selection (FS) module. Specifically, CNN is employed to extract superficial information. Transformer is utilized for exploiting abstract semantic affinity relations. FS module is introduced to reveal discriminative features. Comprehensive experiments demonstrate that can achieve state-of-the-art performance on the CUB-200-2011 dataset (91.0%) and the NABirds dataset (90.9%). © 2022 IEEE

AB - Fine-grained Bird Image Recognition dedicates to achieve accurate bird image classification and it is also a fundamental task in robot vision tracking. Given that endangered bird surveillance and conservation is of great significance for bird protection from extinction, automated approaches are in need to facilitate bird surveillance. In this work, we propose a novel robot vision tracking based method for bird surveillance with an affinity relation-aware model named TBNet that combines CNN and Transformer architecture and has a novel feature selection (FS) module. Specifically, CNN is employed to extract superficial information. Transformer is utilized for exploiting abstract semantic affinity relations. FS module is introduced to reveal discriminative features. Comprehensive experiments demonstrate that can achieve state-of-the-art performance on the CUB-200-2011 dataset (91.0%) and the NABirds dataset (90.9%). © 2022 IEEE

U2 - 10.1109/ROBIO55434.2022.10011861

DO - 10.1109/ROBIO55434.2022.10011861

M3 - RGC 32 - Refereed conference paper (with host publication)

SN - 978-1-6654-8110-6

SP - 662

EP - 667

BT - 2022 IEEE International Conference on Robotics and Biomimetics (ROBIO)

PB - IEEE

T2 - 2022 IEEE International Conference on Robotics and Biomimetics (ROBIO 2022)

Y2 - 5 December 2022 through 9 December 2022

ER -

Affinity Relation-aware Fine-grained Bird Image Recognition for Robot Vision Tracking via Transformers

Abstract

Conference

Funding

UN SDGs

Access Output

Other Access Options

Permanent Link

Fingerprint

Projects

GRF: Event-Based 3D Human Pose Estimation and Tracking

GRF: Gaze Tracking and its Integration with Human-Robot Cooperation

Cite this