Stacked Ensemble of Convolutional Neural Networks for Follicles Detection on Scalp Images

Dymitr Ruta; Ling Cen; Andrzej Ruta; Quang Hieu Vu

doi:10.1007/978-3-031-23480-4_4

Stacked Ensemble of Convolutional Neural Networks for Follicles Detection on Scalp Images

Dymitr Ruta, Ling Cen, Andrzej Ruta, Quang Hieu Vu

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

2 Scopus citations

Abstract

An average person’s head is covered with up to 100000 hairs growing out of follicular openings on the scalp’s skin. Automated hair therapy requires precise detection and localization of follicles on the scalp and still poses a significant challenge for the computer vision and pattern recognition systems. We have proposed an automated vision system for follicles detection based on the classification of digitized microscopic scalp images using an ensemble of convolutional neural networks (CNN). A pool of adapted state-of-the-art CNNs have been transfer-trained on over 700k microscopic skin image regions of 120 × 120 pixels and their outputs further fed to the final stacked ensemble learning layer to capture a wider context of the connected neighboring regions of the original FullHD scalp images. A high validated f1 score (0.7) of detecting regions with follicles beats the industry’s benchmark and brings this technology a step closer towards automated hair treatment as well as other emerging applications such as personal identification based on follicular scalp map.

Original language	British English
Title of host publication	Artificial Intelligence and Soft Computing - 21st International Conference, ICAISC 2022, Proceedings
Editors	Leszek Rutkowski, Leszek Rutkowski, Rafał Scherer, Marcin Korytkowski, Witold Pedrycz, Ryszard Tadeusiewicz, Jacek M. Zurada
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	49-58
Number of pages	10
ISBN (Print)	9783031234798
DOIs	https://doi.org/10.1007/978-3-031-23480-4_4
State	Published - 2023
Event	21st International Conference on Artificial Intelligence and Soft Computing, ICAISC 2022 - Zakopane, Poland Duration: 19 Jun 2022 → 23 Jun 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13589 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	21st International Conference on Artificial Intelligence and Soft Computing, ICAISC 2022
Country/Territory	Poland
City	Zakopane
Period	19/06/22 → 23/06/22

Keywords

Convolutional neural networks
Pattern recognition
Stacked ensemble learning

Access to Document

10.1007/978-3-031-23480-4_4

Cite this

Ruta, D., Cen, L., Ruta, A., & Vu, Q. H. (2023). Stacked Ensemble of Convolutional Neural Networks for Follicles Detection on Scalp Images. In L. Rutkowski, L. Rutkowski, R. Scherer, M. Korytkowski, W. Pedrycz, R. Tadeusiewicz, & J. M. Zurada (Eds.), Artificial Intelligence and Soft Computing - 21st International Conference, ICAISC 2022, Proceedings (pp. 49-58). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13589 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-23480-4_4

Ruta, Dymitr ; Cen, Ling ; Ruta, Andrzej et al. / Stacked Ensemble of Convolutional Neural Networks for Follicles Detection on Scalp Images. Artificial Intelligence and Soft Computing - 21st International Conference, ICAISC 2022, Proceedings. editor / Leszek Rutkowski ; Leszek Rutkowski ; Rafał Scherer ; Marcin Korytkowski ; Witold Pedrycz ; Ryszard Tadeusiewicz ; Jacek M. Zurada. Springer Science and Business Media Deutschland GmbH, 2023. pp. 49-58 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{b3ba9809d3d940bda97feb3507b400c9,

title = "Stacked Ensemble of Convolutional Neural Networks for Follicles Detection on Scalp Images",

abstract = "An average person{\textquoteright}s head is covered with up to 100000 hairs growing out of follicular openings on the scalp{\textquoteright}s skin. Automated hair therapy requires precise detection and localization of follicles on the scalp and still poses a significant challenge for the computer vision and pattern recognition systems. We have proposed an automated vision system for follicles detection based on the classification of digitized microscopic scalp images using an ensemble of convolutional neural networks (CNN). A pool of adapted state-of-the-art CNNs have been transfer-trained on over 700k microscopic skin image regions of 120 × 120 pixels and their outputs further fed to the final stacked ensemble learning layer to capture a wider context of the connected neighboring regions of the original FullHD scalp images. A high validated f1 score (0.7) of detecting regions with follicles beats the industry{\textquoteright}s benchmark and brings this technology a step closer towards automated hair treatment as well as other emerging applications such as personal identification based on follicular scalp map.",

keywords = "Convolutional neural networks, Pattern recognition, Stacked ensemble learning",

author = "Dymitr Ruta and Ling Cen and Andrzej Ruta and Vu, {Quang Hieu}",

note = "Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 21st International Conference on Artificial Intelligence and Soft Computing, ICAISC 2022 ; Conference date: 19-06-2022 Through 23-06-2022",

year = "2023",

doi = "10.1007/978-3-031-23480-4_4",

language = "British English",

isbn = "9783031234798",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

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

pages = "49--58",

editor = "Leszek Rutkowski and Leszek Rutkowski and Rafa{\l} Scherer and Marcin Korytkowski and Witold Pedrycz and Ryszard Tadeusiewicz and Zurada, {Jacek M.}",

booktitle = "Artificial Intelligence and Soft Computing - 21st International Conference, ICAISC 2022, Proceedings",

address = "Germany",

}

Ruta, D, Cen, L, Ruta, A & Vu, QH 2023, Stacked Ensemble of Convolutional Neural Networks for Follicles Detection on Scalp Images. in L Rutkowski, L Rutkowski, R Scherer, M Korytkowski, W Pedrycz, R Tadeusiewicz & JM Zurada (eds), Artificial Intelligence and Soft Computing - 21st International Conference, ICAISC 2022, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13589 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 49-58, 21st International Conference on Artificial Intelligence and Soft Computing, ICAISC 2022, Zakopane, Poland, 19/06/22. https://doi.org/10.1007/978-3-031-23480-4_4

Stacked Ensemble of Convolutional Neural Networks for Follicles Detection on Scalp Images. / Ruta, Dymitr; Cen, Ling; Ruta, Andrzej et al.
Artificial Intelligence and Soft Computing - 21st International Conference, ICAISC 2022, Proceedings. ed. / Leszek Rutkowski; Leszek Rutkowski; Rafał Scherer; Marcin Korytkowski; Witold Pedrycz; Ryszard Tadeusiewicz; Jacek M. Zurada. Springer Science and Business Media Deutschland GmbH, 2023. p. 49-58 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13589 LNAI).

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

TY - GEN

T1 - Stacked Ensemble of Convolutional Neural Networks for Follicles Detection on Scalp Images

AU - Ruta, Dymitr

AU - Cen, Ling

AU - Ruta, Andrzej

AU - Vu, Quang Hieu

PY - 2023

Y1 - 2023

N2 - An average person’s head is covered with up to 100000 hairs growing out of follicular openings on the scalp’s skin. Automated hair therapy requires precise detection and localization of follicles on the scalp and still poses a significant challenge for the computer vision and pattern recognition systems. We have proposed an automated vision system for follicles detection based on the classification of digitized microscopic scalp images using an ensemble of convolutional neural networks (CNN). A pool of adapted state-of-the-art CNNs have been transfer-trained on over 700k microscopic skin image regions of 120 × 120 pixels and their outputs further fed to the final stacked ensemble learning layer to capture a wider context of the connected neighboring regions of the original FullHD scalp images. A high validated f1 score (0.7) of detecting regions with follicles beats the industry’s benchmark and brings this technology a step closer towards automated hair treatment as well as other emerging applications such as personal identification based on follicular scalp map.

AB - An average person’s head is covered with up to 100000 hairs growing out of follicular openings on the scalp’s skin. Automated hair therapy requires precise detection and localization of follicles on the scalp and still poses a significant challenge for the computer vision and pattern recognition systems. We have proposed an automated vision system for follicles detection based on the classification of digitized microscopic scalp images using an ensemble of convolutional neural networks (CNN). A pool of adapted state-of-the-art CNNs have been transfer-trained on over 700k microscopic skin image regions of 120 × 120 pixels and their outputs further fed to the final stacked ensemble learning layer to capture a wider context of the connected neighboring regions of the original FullHD scalp images. A high validated f1 score (0.7) of detecting regions with follicles beats the industry’s benchmark and brings this technology a step closer towards automated hair treatment as well as other emerging applications such as personal identification based on follicular scalp map.

KW - Convolutional neural networks

KW - Pattern recognition

KW - Stacked ensemble learning

UR - http://www.scopus.com/inward/record.url?scp=85149689643&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-23480-4_4

DO - 10.1007/978-3-031-23480-4_4

M3 - Conference contribution

AN - SCOPUS:85149689643

SN - 9783031234798

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 49

EP - 58

BT - Artificial Intelligence and Soft Computing - 21st International Conference, ICAISC 2022, Proceedings

A2 - Rutkowski, Leszek

A2 - Scherer, Rafał

A2 - Korytkowski, Marcin

A2 - Pedrycz, Witold

A2 - Tadeusiewicz, Ryszard

A2 - Zurada, Jacek M.

PB - Springer Science and Business Media Deutschland GmbH

T2 - 21st International Conference on Artificial Intelligence and Soft Computing, ICAISC 2022

Y2 - 19 June 2022 through 23 June 2022

ER -

Ruta D, Cen L, Ruta A, Vu QH. Stacked Ensemble of Convolutional Neural Networks for Follicles Detection on Scalp Images. In Rutkowski L, Rutkowski L, Scherer R, Korytkowski M, Pedrycz W, Tadeusiewicz R, Zurada JM, editors, Artificial Intelligence and Soft Computing - 21st International Conference, ICAISC 2022, Proceedings. Springer Science and Business Media Deutschland GmbH. 2023. p. 49-58. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-23480-4_4