Abstract

Detections of material regions on CT-scans of solids are commonly treated manually by an expert. Although such manual detections have many advantages, some amount of human error is also incorporated. Moreover, expert opinions may vary significantly. We present an application of the k-means++ clustering as an alternative option to manual way of material area detections. k-means++ clustering is derived from k-means (the method of vector quantization, originally from signal processing), popular for cluster analysis in data mining and image processing communities. The algorithm s main advantages are its simple implementation and fast convergence to a local optimum of an objective function. We benchmark the suggested approach on transverse CT-scans of a fibre-reinforced concrete solid. Moreover, we introduce a technique for processing air distribution, such that the appropriate pixels detected as the pixels of air are converted into pixels representing concrete. The technique is based on the connected component algorithm. Benchmark and results of proposed method conclude the paper.

Highlights

  • Image processing techniques form an essential part of modern simulation methods especially in preprocessing, post-processing and visualizations

  • The Region Of Interest ΩROI ⊂ Ω is a part of an image function area which is selected by the segmentator/classifier

  • For extracting the ROI from an image function area Ω, a binary mask is used in image processing field

Read more

Summary

Fundamental Notations and Definitions

Let Ω ⊂ R2 be the image function area. It is suitable to define the image function f (x, y) as the mapping f : Ω → 0, 255 in case of a CT-scan image. The Region Of Interest ΩROI ⊂ Ω is a part of an image function area which is selected by the segmentator/classifier (a program or human user). Region ΩROI is typically a part of image foreground and Ω \ ΩROI is interpreted as the background of the image

Introduction
The k -means Algorithm
Non-Local Material
Conclusions
Future Work
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call