Multi-Scale Facial Scanning via Spatial Lstm for Latent Facial Feature Representation

Seong Tae Kim; Yeoreum Choi; Yong Man Ro

doi:10.23919/BIOSIG.2017.8053515

Multi-Scale Facial Scanning via Spatial Lstm for Latent Facial Feature Representation

Seong Tae Kim
, Yeoreum Choi
, Yong Man Ro

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

Abstract

In the past few decades, automatic face recognition has been an important vision task. In this paper, we exploit the spatial relationships of facial local regions by using a novel deep network. In the proposed method, face is spatially scanned with spatial long short-term memory (LSTM) to encode the spatial correlation of facial regions. Moreover, with facial regions of various scales, the complementary information of the multi-scale facial features is encoded. Experimental results on public database showed that the proposed method outperformed the conventional methods by improving the face recognition accuracy under illumination variation.

Original language	English
Title of host publication	2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017
Editors	Arslan Bromme, Christoph Busch, Antitza Dantcheva, Christian Rathgeb, Andreas Uhl
Publisher	Gesellschaft fur Informatik (GI)
ISBN (Electronic)	9783885796640
DOIs	https://doi.org/10.23919/BIOSIG.2017.8053515
Publication status	Published - 28 Sept 2017
Event	2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017 - Darmstadt, Germany Duration: 20 Sept 2017 → 22 Sept 2017

Publication series

Name	Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI)
Volume	0
ISSN (Print)	1617-5468
ISSN (Electronic)	2944-7682

Conference

Conference	2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017
Country/Territory	Germany
City	Darmstadt
Period	20/09/17 → 22/09/17

Bibliographical note

Publisher Copyright:
© 2017 Gesellschaft fuer Informatik.

Keywords

Face recognition
deep learning
facial feature representation
spatial LSTM

Access to Document

10.23919/BIOSIG.2017.8053515

Cite this

Kim, S. T., Choi, Y., & Ro, Y. M. (2017). Multi-Scale Facial Scanning via Spatial Lstm for Latent Facial Feature Representation. In A. Bromme, C. Busch, A. Dantcheva, C. Rathgeb, & A. Uhl (Eds.), 2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017 Article 8053515 (Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI); Vol. 0). Gesellschaft fur Informatik (GI). https://doi.org/10.23919/BIOSIG.2017.8053515

Kim, Seong Tae ; Choi, Yeoreum ; Ro, Yong Man. / Multi-Scale Facial Scanning via Spatial Lstm for Latent Facial Feature Representation. 2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017. editor / Arslan Bromme ; Christoph Busch ; Antitza Dantcheva ; Christian Rathgeb ; Andreas Uhl. Gesellschaft fur Informatik (GI), 2017. (Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI)).

@inproceedings{d57676e787da4f05880764254bc8160b,

title = "Multi-Scale Facial Scanning via Spatial Lstm for Latent Facial Feature Representation",

abstract = "In the past few decades, automatic face recognition has been an important vision task. In this paper, we exploit the spatial relationships of facial local regions by using a novel deep network. In the proposed method, face is spatially scanned with spatial long short-term memory (LSTM) to encode the spatial correlation of facial regions. Moreover, with facial regions of various scales, the complementary information of the multi-scale facial features is encoded. Experimental results on public database showed that the proposed method outperformed the conventional methods by improving the face recognition accuracy under illumination variation.",

keywords = "Face recognition, deep learning, facial feature representation, spatial LSTM",

author = "Kim, \{Seong Tae\} and Yeoreum Choi and Ro, \{Yong Man\}",

note = "Publisher Copyright: {\textcopyright} 2017 Gesellschaft fuer Informatik.; 2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017 ; Conference date: 20-09-2017 Through 22-09-2017",

year = "2017",

month = sep,

day = "28",

doi = "10.23919/BIOSIG.2017.8053515",

language = "English",

series = "Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI)",

publisher = "Gesellschaft fur Informatik (GI)",

editor = "Arslan Bromme and Christoph Busch and Antitza Dantcheva and Christian Rathgeb and Andreas Uhl",

booktitle = "2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017",

address = "Germany",

}

Kim, ST, Choi, Y & Ro, YM 2017, Multi-Scale Facial Scanning via Spatial Lstm for Latent Facial Feature Representation. in A Bromme, C Busch, A Dantcheva, C Rathgeb & A Uhl (eds), 2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017., 8053515, Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI), vol. 0, Gesellschaft fur Informatik (GI), 2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017, Darmstadt, Germany, 20/09/17. https://doi.org/10.23919/BIOSIG.2017.8053515

Multi-Scale Facial Scanning via Spatial Lstm for Latent Facial Feature Representation. / Kim, Seong Tae; Choi, Yeoreum; Ro, Yong Man.
2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017. ed. / Arslan Bromme; Christoph Busch; Antitza Dantcheva; Christian Rathgeb; Andreas Uhl. Gesellschaft fur Informatik (GI), 2017. 8053515 (Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI); Vol. 0).

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

TY - GEN

T1 - Multi-Scale Facial Scanning via Spatial Lstm for Latent Facial Feature Representation

AU - Kim, Seong Tae

AU - Choi, Yeoreum

AU - Ro, Yong Man

PY - 2017/9/28

Y1 - 2017/9/28

N2 - In the past few decades, automatic face recognition has been an important vision task. In this paper, we exploit the spatial relationships of facial local regions by using a novel deep network. In the proposed method, face is spatially scanned with spatial long short-term memory (LSTM) to encode the spatial correlation of facial regions. Moreover, with facial regions of various scales, the complementary information of the multi-scale facial features is encoded. Experimental results on public database showed that the proposed method outperformed the conventional methods by improving the face recognition accuracy under illumination variation.

AB - In the past few decades, automatic face recognition has been an important vision task. In this paper, we exploit the spatial relationships of facial local regions by using a novel deep network. In the proposed method, face is spatially scanned with spatial long short-term memory (LSTM) to encode the spatial correlation of facial regions. Moreover, with facial regions of various scales, the complementary information of the multi-scale facial features is encoded. Experimental results on public database showed that the proposed method outperformed the conventional methods by improving the face recognition accuracy under illumination variation.

KW - Face recognition

KW - deep learning

KW - facial feature representation

KW - spatial LSTM

UR - https://www.scopus.com/pages/publications/85034589197

U2 - 10.23919/BIOSIG.2017.8053515

DO - 10.23919/BIOSIG.2017.8053515

M3 - Conference contribution

AN - SCOPUS:85034589197

T3 - Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI)

BT - 2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017

A2 - Bromme, Arslan

A2 - Busch, Christoph

A2 - Dantcheva, Antitza

A2 - Rathgeb, Christian

A2 - Uhl, Andreas

PB - Gesellschaft fur Informatik (GI)

T2 - 2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017

Y2 - 20 September 2017 through 22 September 2017

ER -

Kim ST, Choi Y, Ro YM. Multi-Scale Facial Scanning via Spatial Lstm for Latent Facial Feature Representation. In Bromme A, Busch C, Dantcheva A, Rathgeb C, Uhl A, editors, 2017 International Conference of the Biometrics Special Interest Group, BIOSIG 2017. Gesellschaft fur Informatik (GI). 2017. 8053515. (Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI)). doi: 10.23919/BIOSIG.2017.8053515

Multi-Scale Facial Scanning via Spatial Lstm for Latent Facial Feature Representation

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this