Skip to main content

METHODS article

Front. Plant Sci.
Sec. Technical Advances in Plant Science
Volume 15 - 2024 | doi: 10.3389/fpls.2024.1414849

Comparative Analysis of Stomata Pore Instance Segmentation: Mask R-CNN vs. YOLOv8 on PhenomicsStomata Dataset

Provisionally accepted
Thanh Tuan Thai Thanh Tuan Thai 1,2,3Kibon Ku Kibon Ku 1Anh Tuan Le Anh Tuan Le 2,4San Su Min Oh San Su Min Oh 1Ngo Hoang Phan Ngo Hoang Phan 2,4Yong Suk Chung Yong Suk Chung 1,5*
  • 1 Jeju National University, Jeju City, Jeju, Republic of Korea
  • 2 Vietnam National University, Ho Chi Minh City, Ho Chi Minh City, Vietnam
  • 3 Ho Chi Minh City University of Information Technology, Ho Chi Minh City, Vietnam
  • 4 Ho Chi Minh City University of Science, Ho Chi Minh City, Vietnam
  • 5 Phytomix Corporation, Jeju 63022, Republic of Korea

The final, formatted version of the article will be published soon.

    This study conducts a rigorous comparative analysis between two cutting-edge instance segmentation methods, Mask R-CNN and YOLOv8, focusing on stomata pore analysis. A novel dataset specifically tailored for stomata pore instance segmentation, named PhenomicsStomata, was introduced. This dataset posed challenges such as low resolution and image imperfections, prompting the application of advanced preprocessing techniques, including image enhancement using the Lucy-Richardson Algorithm. The models underwent comprehensive evaluation, considering accuracy, precision, and recall as key parameters. Notably, YOLOv8 demonstrated superior performance over Mask R-CNN, particularly in accurately calculating stomata pore dimensions. Beyond this comparative study, the implications of our findings extend across diverse biological research, providing a robust foundation for advancing our understanding of plant physiology. Furthermore, the preprocessing enhancements offer valuable insights for refining image analysis techniques, showcasing the potential for broader applications in scientific domains. This research marks a significant stride in unraveling the complexities of plant structures, offering both theoretical insights and practical applications in scientific research.

    Keywords: stomata, phenotyping, Instance segmentation, Mask-RCNN, YOLO

    Received: 09 Apr 2024; Accepted: 14 Oct 2024.

    Copyright: © 2024 Thai, Ku, Le, Oh, Phan and Chung. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

    * Correspondence: Yong Suk Chung, Jeju National University, Jeju City, 690-756, Jeju, Republic of Korea

    Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.