Variance Orientation Transform Detection of Early Osteoarthritis in Knee Trabecular Bone

Stefan Ovidiu Oancea
Variance Orientation Transform Detection of Early Osteoarthritis in Knee Trabecular Bone
[Thesis]

Information

Abstract

Since the fractal properties of the knee trabecular bone were discovered, fractal methods for analyzing bone surface radiographic projections have gained more attention. This is partly due to the fact that radiography is the cheapest imaging technique in routine clinical screening and partly due to the fact that it was shown that the trabecular bones of osteoarthritic patients indicate early deformations, even long before the characteristic join loss occurs. The ultimate goal of such an algorithm would be to differentiate healthy from unhealthy trabecular bone.

This paper presents a report of our implementation of the Variance Orientation Transform (VOT) algorithm, a fractal method, which unlike other similar methods, is able to quantify bone texture in different directions and over different scales of measurement.

It is based on the idea that a single fractal dimension value is not enough to describe such a complex structure as the trabecular bone and thus, VOT calculates more descriptive fractal dimensions called fractal signatures (FSs).

In Chapters 1 and 2 we introduce the notion of fractals and the theoretical background behind them and the VOT algorithm. In Chapter 3 similar techniques for analyzing trabecular bone are presented and in Chapter 4 our particular attempt at implementing VOT is described in detail; moreover, in the same Chapter VOT is validated using some artificially generated fractal surfaces and the ability of differentiating healthy and affected bone is also investigated. The last Chapter, Chapter 5, covers further possible ideas of improving and testing of the algorithm.

Additional Files and Images

Weblinks

No further information available.

BibTeX

@bachelorsthesis{Oancea_Stefan_2016_VOT,
  title =      "Variance Orientation Transform Detection of Early
               Osteoarthritis in Knee Trabecular Bone",
  author =     "Stefan Ovidiu Oancea",
  year =       "2016",
  abstract =   "Since the fractal properties of the knee trabecular bone
               were discovered, fractal methods for analyzing bone surface
               radiographic projections have gained more attention. This is
               partly due to the fact that radiography is the cheapest
               imaging technique in routine clinical screening and partly
               due to the fact that it was shown that the trabecular bones
               of osteoarthritic patients indicate early deformations, even
               long before the  characteristic join loss occurs. The
               ultimate goal of such an algorithm would be to differentiate
               healthy from unhealthy trabecular bone.  This paper presents
               a report of our implementation of the Variance Orientation
               Transform (VOT) algorithm, a fractal method, which unlike
               other similar methods, is able to quantify bone texture in
               different directions and over different scales of
               measurement.  It is based on the idea that a single fractal
               dimension value is not enough to describe such a complex
               structure as the trabecular bone and thus, VOT calculates
               more descriptive fractal dimensions called fractal
               signatures (FSs).  In Chapters 1 and 2 we introduce the
               notion of fractals and the theoretical background behind
               them and the VOT algorithm. In Chapter 3 similar techniques
               for analyzing trabecular bone are presented and in Chapter 4
               our particular attempt at implementing VOT is described in
               detail; moreover, in the same Chapter VOT is validated using
               some artificially generated fractal surfaces and the ability
               of differentiating healthy and affected bone is also
               investigated. The last Chapter, Chapter 5, covers further
               possible ideas of improving and testing of the algorithm.",
  month =      mar,
  address =    "Favoritenstrasse 9-11/186, A-1040 Vienna, Austria",
  school =     "Institute of Computer Graphics and Algorithms, Vienna
               University of Technology",
  URL =        "https://www.cg.tuwien.ac.at/research/publications/2016/Oancea_Stefan_2016_VOT/",
}