About this Abstract |
| Meeting |
2026 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2026)
|
| Symposium
|
2026 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2026)
|
| Presentation Title |
Automated Fiber Segmentation and Characterization in Ceramic Composite Micrographs1 |
| Author(s) |
Andi Scarola, Jason McCleary, Carmen Rocha Gutierrez, Rosario Lara, Francisco Medina |
| On-Site Speaker (Planned) |
Andi Scarola |
| Abstract Scope |
Fiber characterization in ceramic composite micrographs is important for understanding microstructural organization, but manual analysis is time-intensive, subjective, and challenging to scale. In addition, traditional semi-automated workflows such as ImageJ can become difficult to tune reliably for large micrographs with many features and complex background structures. This work presents an automated pipeline for fiber segmentation and characterization using overlapping tiled image processing, a U-Net-based model for fiber detection, and downstream analysis for estimating fiber density and orientation. Rather than shrinking the full micrograph before analysis, the system divides the image into overlapping tiles, so fine details can be preserved. During analysis, each tile is segmented separately, and the tile-level results are then combined back into a full-image output for downstream measurements such as fiber density and orientation. Evaluation is based on comparison with manual and ImageJ-based analysis to assess reproducibility in ceramic composite imaging. |
| Proceedings Inclusion? |
Undecided |