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 proceedingConference contributionpeer-review

    1 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 languageBritish English
    Title of host publicationArtificial Intelligence and Soft Computing - 21st International Conference, ICAISC 2022, Proceedings
    EditorsLeszek Rutkowski, Leszek Rutkowski, Rafał Scherer, Marcin Korytkowski, Witold Pedrycz, Ryszard Tadeusiewicz, Jacek M. Zurada
    PublisherSpringer Science and Business Media Deutschland GmbH
    Pages49-58
    Number of pages10
    ISBN (Print)9783031234798
    DOIs
    StatePublished - 2023
    Event21st International Conference on Artificial Intelligence and Soft Computing, ICAISC 2022 - Zakopane, Poland
    Duration: 19 Jun 202223 Jun 2022

    Publication series

    NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    Volume13589 LNAI
    ISSN (Print)0302-9743
    ISSN (Electronic)1611-3349

    Conference

    Conference21st International Conference on Artificial Intelligence and Soft Computing, ICAISC 2022
    Country/TerritoryPoland
    CityZakopane
    Period19/06/2223/06/22

    Keywords

    • Convolutional neural networks
    • Pattern recognition
    • Stacked ensemble learning

    Fingerprint

    Dive into the research topics of 'Stacked Ensemble of Convolutional Neural Networks for Follicles Detection on Scalp Images'. Together they form a unique fingerprint.

    Cite this