Related Experiment Video
Updated: Aug 14, 2026
12:08
From Voxels to Knowledge: A Practical Guide to the Segmentation of Complex Electron Microscopy 3D-Data
Published on: August 13, 2014
FastSAM-Based Automated Segmentation and Data Extraction for Pore Structures of Foamed Concrete
Luchang Xiong1,2, Siyu Du2, Zhijun Wan2
1School of Safety Engineering, China University of Mining and Technology, Xuzhou 221116, China.
Materials (Basel, Switzerland)
|August 13, 2026
Summary
A new FastSAM-DP workflow accurately segments foamed concrete pores and quantifies their structure. This automated method improves pore-structure assessment for reliable material characterization.
Area of Science:
- Materials Science
- Image Analysis
- Civil Engineering
Background:
- Accurate pore-structure recognition in foamed concrete is difficult due to blurred pore walls, interconnected pores, and scale-dependent image statistics.
- Existing methods struggle with reliable segmentation and quantitative descriptor extraction from micrographs.
Purpose of the Study:
- To develop and validate an automated workflow (FastSAM-DP) for precise pore segmentation and comprehensive pore-structure assessment in foamed concrete micrographs.
- To extract key pore descriptors including size coefficient of variation (CV), circularity (Ci), number density (N), uniformity index (UI), and large-pore area fraction (FL).
Main Methods:
- The FastSAM-DP workflow combines FastSAM for instance segmentation with the Douglas-Peucker (DP) algorithm for contour regularization and simplification.
- Pore instance masks are generated, followed by geometric simplification to extract quantitative pore-structure descriptors.
- Workflow configuration was optimized using development data and validated through rigorous internal evaluations on held-out datasets.
Main Results:
- FastSAM-DP achieved a high instance F1 score of 0.732, significantly outperforming the fixed Otsu-Watershed method (0.050) on a complete held-out set.
- The workflow demonstrated strong agreement for the pore-structure quality index (PSQI) with a correlation coefficient (r) of 0.828.
- Validation on a sensitivity subset yielded an instance F1 of 0.752 and PSQI agreement of r = 0.838, indicating robustness.
Conclusions:
- The FastSAM-DP workflow provides an effective and automated solution for pore segmentation and structure assessment in foamed concrete.
- The method enables reliable batch screening and comparative analysis of pore structures at consistent magnifications.
- This advancement facilitates more accurate material characterization and quality control in concrete applications.
