Feature Redirection Network for Few-Shot Classification

Yanan Wang; Guoqiang Zhong; Yuxu Mao; Kaizhu Huang

doi:10.1007/978-3-030-63820-7_48

Feature Redirection Network for Few-Shot Classification

Yanan Wang, Guoqiang Zhong^*, Yuxu Mao, Kaizhu Huang

^*Corresponding author for this work

School of Advanced Technology

Ocean University of China

Research output: Chapter in Book or Report/Conference proceeding › Conference Proceeding › peer-review

2 Citations (Scopus)

Abstract

Few-shot classification aims to learn novel categories by giving few labeled samples. How to make best use of the limited data to obtain a learner with fast learning ability has become a challenging problem. In this paper, we propose a feature redirection network (FRNet) for few-shot classification to make the features more discriminative. The proposed FRNet not only highlights relevant category features of support samples, but also learns how to generate task-relevant features of query samples. Experiments conducted on three datasets have demonstrate its superiority over the state-of-the-art methods.

Original language	English
Title of host publication	Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings
Editors	Haiqin Yang, Kitsuchart Pasupa, Andrew Chi-Sing Leung, James T. Kwok, Jonathan H. Chan, Irwin King
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	418-425
Number of pages	8
ISBN (Print)	9783030638191
DOIs	https://doi.org/10.1007/978-3-030-63820-7_48
Publication status	Published - 2020
Event	27th International Conference on Neural Information Processing, ICONIP 2020 - Bangkok, Thailand Duration: 18 Nov 2020 → 22 Nov 2020

Publication series

Name	Communications in Computer and Information Science
Volume	1332
ISSN (Print)	1865-0929
ISSN (Electronic)	1865-0937

Conference

Conference	27th International Conference on Neural Information Processing, ICONIP 2020
Country/Territory	Thailand
City	Bangkok
Period	18/11/20 → 22/11/20

Keywords

Feature redirection
Few-shot classification
Task-relevant

Access to Document

10.1007/978-3-030-63820-7_48

Cite this

Wang, Y., Zhong, G., Mao, Y., & Huang, K. (2020). Feature Redirection Network for Few-Shot Classification. In H. Yang, K. Pasupa, A. C.-S. Leung, J. T. Kwok, J. H. Chan, & I. King (Eds.), Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings (pp. 418-425). (Communications in Computer and Information Science; Vol. 1332). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-63820-7_48

Wang, Yanan ; Zhong, Guoqiang ; Mao, Yuxu et al. / Feature Redirection Network for Few-Shot Classification. Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings. editor / Haiqin Yang ; Kitsuchart Pasupa ; Andrew Chi-Sing Leung ; James T. Kwok ; Jonathan H. Chan ; Irwin King. Springer Science and Business Media Deutschland GmbH, 2020. pp. 418-425 (Communications in Computer and Information Science).

@inproceedings{d5091fa9d01f46dca622bd5cbe3b22a8,

title = "Feature Redirection Network for Few-Shot Classification",

abstract = "Few-shot classification aims to learn novel categories by giving few labeled samples. How to make best use of the limited data to obtain a learner with fast learning ability has become a challenging problem. In this paper, we propose a feature redirection network (FRNet) for few-shot classification to make the features more discriminative. The proposed FRNet not only highlights relevant category features of support samples, but also learns how to generate task-relevant features of query samples. Experiments conducted on three datasets have demonstrate its superiority over the state-of-the-art methods.",

keywords = "Feature redirection, Few-shot classification, Task-relevant",

author = "Yanan Wang and Guoqiang Zhong and Yuxu Mao and Kaizhu Huang",

note = "Funding Information: Acknowledgments. This work was supported by the Major Project for New Generation of AI under Grant No. 2018AAA0100400, the National Natural Science Foundation of China (NSFC) under Grants No. 41706010 and No. 61876155, the Joint Fund of the Equipments Pre-Research and Ministry of Education of China under Grant No. 6141A020337, the Natural Science Foundation of Jiangsu Province under Grant No. BK20181189, the Key Program Special Fund in XJTLU under Grants No. KSF-A-01, KSF-T-06, KSF-E-26, KSF-P-02 and KSF-A-10, and the Fundamental Research Funds for the Central Universities of China. Publisher Copyright: {\textcopyright} 2020, Springer Nature Switzerland AG.; 27th International Conference on Neural Information Processing, ICONIP 2020 ; Conference date: 18-11-2020 Through 22-11-2020",

year = "2020",

doi = "10.1007/978-3-030-63820-7_48",

language = "English",

isbn = "9783030638191",

series = "Communications in Computer and Information Science",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "418--425",

editor = "Haiqin Yang and Kitsuchart Pasupa and Leung, {Andrew Chi-Sing} and Kwok, {James T.} and Chan, {Jonathan H.} and Irwin King",

booktitle = "Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings",

}

Wang, Y, Zhong, G, Mao, Y & Huang, K 2020, Feature Redirection Network for Few-Shot Classification. in H Yang, K Pasupa, AC-S Leung, JT Kwok, JH Chan & I King (eds), Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings. Communications in Computer and Information Science, vol. 1332, Springer Science and Business Media Deutschland GmbH, pp. 418-425, 27th International Conference on Neural Information Processing, ICONIP 2020, Bangkok, Thailand, 18/11/20. https://doi.org/10.1007/978-3-030-63820-7_48

Feature Redirection Network for Few-Shot Classification. / Wang, Yanan; Zhong, Guoqiang; Mao, Yuxu et al.
Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings. ed. / Haiqin Yang; Kitsuchart Pasupa; Andrew Chi-Sing Leung; James T. Kwok; Jonathan H. Chan; Irwin King. Springer Science and Business Media Deutschland GmbH, 2020. p. 418-425 (Communications in Computer and Information Science; Vol. 1332).

Research output: Chapter in Book or Report/Conference proceeding › Conference Proceeding › peer-review

TY - GEN

T1 - Feature Redirection Network for Few-Shot Classification

AU - Wang, Yanan

AU - Zhong, Guoqiang

AU - Mao, Yuxu

AU - Huang, Kaizhu

N1 - Funding Information: Acknowledgments. This work was supported by the Major Project for New Generation of AI under Grant No. 2018AAA0100400, the National Natural Science Foundation of China (NSFC) under Grants No. 41706010 and No. 61876155, the Joint Fund of the Equipments Pre-Research and Ministry of Education of China under Grant No. 6141A020337, the Natural Science Foundation of Jiangsu Province under Grant No. BK20181189, the Key Program Special Fund in XJTLU under Grants No. KSF-A-01, KSF-T-06, KSF-E-26, KSF-P-02 and KSF-A-10, and the Fundamental Research Funds for the Central Universities of China. Publisher Copyright: © 2020, Springer Nature Switzerland AG.

PY - 2020

Y1 - 2020

N2 - Few-shot classification aims to learn novel categories by giving few labeled samples. How to make best use of the limited data to obtain a learner with fast learning ability has become a challenging problem. In this paper, we propose a feature redirection network (FRNet) for few-shot classification to make the features more discriminative. The proposed FRNet not only highlights relevant category features of support samples, but also learns how to generate task-relevant features of query samples. Experiments conducted on three datasets have demonstrate its superiority over the state-of-the-art methods.

AB - Few-shot classification aims to learn novel categories by giving few labeled samples. How to make best use of the limited data to obtain a learner with fast learning ability has become a challenging problem. In this paper, we propose a feature redirection network (FRNet) for few-shot classification to make the features more discriminative. The proposed FRNet not only highlights relevant category features of support samples, but also learns how to generate task-relevant features of query samples. Experiments conducted on three datasets have demonstrate its superiority over the state-of-the-art methods.

KW - Feature redirection

KW - Few-shot classification

KW - Task-relevant

UR - http://www.scopus.com/inward/record.url?scp=85097290102&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-63820-7_48

DO - 10.1007/978-3-030-63820-7_48

M3 - Conference Proceeding

AN - SCOPUS:85097290102

SN - 9783030638191

T3 - Communications in Computer and Information Science

SP - 418

EP - 425

BT - Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings

A2 - Yang, Haiqin

A2 - Pasupa, Kitsuchart

A2 - Leung, Andrew Chi-Sing

A2 - Kwok, James T.

A2 - Chan, Jonathan H.

A2 - King, Irwin

PB - Springer Science and Business Media Deutschland GmbH

T2 - 27th International Conference on Neural Information Processing, ICONIP 2020

Y2 - 18 November 2020 through 22 November 2020

ER -

Wang Y, Zhong G, Mao Y, Huang K. Feature Redirection Network for Few-Shot Classification. In Yang H, Pasupa K, Leung ACS, Kwok JT, Chan JH, King I, editors, Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings. Springer Science and Business Media Deutschland GmbH. 2020. p. 418-425. (Communications in Computer and Information Science). doi: 10.1007/978-3-030-63820-7_48

Feature Redirection Network for Few-Shot Classification

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this