Benchmarking and sensitivity analysis of segmentation methods for image-based fiber detection in composites
Abstract
The accurate characterization of the microstructure of fiber-reinforced polymer composites is crucial for quality control in manufacturing, material design, and robust performance prediction. Most of the characterization methods rely on the analysis of 2D or 3D images. However, the composites community lacks commonly shared best practices and a clear quantitative understanding of how different image processing approaches compare. This work presents a benchmarking exercise on image processing of fiber-reinforced composite materials to address this challenge. Processing three cross-section micrographs, acquired from a single unidirectional composite sample using typical yet distinct imaging protocols, 11 participants from 8 research institutions extracted fiber centroids and radii. The estimated fiber volume fraction (Vf) values for a single cross-section varied between participants from 0.42 to 0.65. The results highlight the sensitivity of different methods to factors like illumination inhomogeneities, pixel density, polishing-related surface defects, and fiber packing. Considering these observations, several best practices for image analysis are identified, providing critical insight into the methods’ sensitivities. To promote transparency and community uptake, the benchmark dataset, evaluation scripts, documentation, and algorithms that were open-sourced are made openly available through a dedicated website, paving the way towards more reliable microstructural analysis in composite materials and ultimately more standardized protocols.
Related
- The conductance filtration: sink-relative persistence and exact edge sensitivity for heat transport in particle-filled composites
- Online-compatible EMG onset detection: benchmarking common detectors under large signal variability
- Benchmarking One-Dimensional Quantum Lattice Boltzmann Methods on Quantum Hardware
- Benchmarking Hierarchical Double-Ended Transition-State Search Methods for Metal-Catalyzed Reactions.
- End-to-end prostate lesion segmentation and detection from undersampled diffusion-weighted MRI
BibTeX
@article{yuksel2026benchmarking,
title = {{Benchmarking and sensitivity analysis of segmentation methods for image-based fiber detection in composites}},
author = {Onur Yuksel and Guillaume Broggi and Robin Hartley and Vincent Maes and Theo Baumard and Christophe Binetruy and Anders Bjorholm Dahl and Christian Breite and Clemens Dransfeld and Wouter Grouve and Silvia Gomarasca and Rui Guo and Adrien Le Reun and Arthur Levy and David May and Mahoor Mehdikhani and Andrea Miene and Lars Pilgaard Mikkelsen and Diwakar Singh and Yentl Swolfs and Elena Syerko and James Kratz and Baris Caglar},
journal = {Composites Part A: Applied Science and Manufacturing},
year = {2026},
month = sep,
doi = {10.1016/j.compositesa.2026.110150},
url = {https://hal.science/hal-05741249v1},
note = {CC BY-NC 4.0},
}
RIS
TY - JOUR
TI - Benchmarking and sensitivity analysis of segmentation methods for image-based fiber detection in composites
AU - Onur Yuksel
AU - Guillaume Broggi
AU - Robin Hartley
AU - Vincent Maes
AU - Theo Baumard
AU - Christophe Binetruy
AU - Anders Bjorholm Dahl
AU - Christian Breite
AU - Clemens Dransfeld
AU - Wouter Grouve
AU - Silvia Gomarasca
AU - Rui Guo
AU - Adrien Le Reun
AU - Arthur Levy
AU - David May
AU - Mahoor Mehdikhani
AU - Andrea Miene
AU - Lars Pilgaard Mikkelsen
AU - Diwakar Singh
AU - Yentl Swolfs
AU - Elena Syerko
AU - James Kratz
AU - Baris Caglar
PY - 2026
DA - 2026/09/08
JO - Composites Part A: Applied Science and Manufacturing
DO - 10.1016/j.compositesa.2026.110150
UR - https://hal.science/hal-05741249v1
AB - The accurate characterization of the microstructure of fiber-reinforced polymer composites is crucial for quality control in manufacturing, material design, and robust performance prediction. Most of the characterization methods rely on the analysis of 2D or 3D images. However, the composites community lacks commonly shared best practices and a clear quantitative understanding of how different image processing approaches compare. This work presents a benchmarking exercise on image processing of fiber-reinforced composite materials to address this challenge. Processing three cross-section micrographs, acquired from a single unidirectional composite sample using typical yet distinct imaging protocols, 11 participants from 8 research institutions extracted fiber centroids and radii. The estimated fiber volume fraction (Vf) values for a single cross-section varied between participants from 0.42 to 0.65. The results highlight the sensitivity of different methods to factors like illumination inhomogeneities, pixel density, polishing-related surface defects, and fiber packing. Considering these observations, several best practices for image analysis are identified, providing critical insight into the methods’ sensitivities. To promote transparency and community uptake, the benchmark dataset, evaluation scripts, documentation, and algorithms that were open-sourced are made openly available through a dedicated website, paving the way towards more reliable microstructural analysis in composite materials and ultimately more standardized protocols.
SN - 1359-835X
ER -