Interactive volumetric segmentation for textile micro-tomography data using wavelets and nonlocal means

J. Michael L. MacNeil, Daniela M. Ushizima, Francesco Panerai, Nagi N. Mansour, Harold S. Barnard, Dilworth Y. Parkinson

Research output: Contribution to journalArticlepeer-review


This work addresses segmentation of volumetric images of woven carbon fiber textiles from micro-tomography data. We propose a semi-supervised algorithm to classify carbon fibers that requires sparse input as opposed to completely labeled images. The main contributions are: (a) design of effective discriminative classifiers, for three-dimensional textile samples, trained on wavelet features for segmentation; (b) coupling of previous step with nonlocal means as simple, efficient alternative to the Potts model; and (c) demonstration of reuse of classifier to diverse samples containing similar content. We evaluate our work by curating test sets of voxels in the absence of a complete ground truth mask. The algorithm obtains an average 0.95 F1 score on test sets and average F1 score of 0.93 on new samples. We conclude with discussion of failure cases and propose future directions toward analysis of spatiotemporal high-resolution micro-tomography images.

Original languageEnglish (US)
Pages (from-to)338-353
Number of pages16
JournalStatistical Analysis and Data Mining
Issue number4
StatePublished - Aug 2019
Externally publishedYes


  • 3D image processing
  • 3D segmentation
  • 3D woven carbon fiber
  • composites
  • machine learning
  • microCT
  • neural networks

ASJC Scopus subject areas

  • Analysis
  • Information Systems
  • Computer Science Applications


Dive into the research topics of 'Interactive volumetric segmentation for textile micro-tomography data using wavelets and nonlocal means'. Together they form a unique fingerprint.

Cite this