Texture Analysis of Diffraction Enhanced Synchrotron Images of Trabecular Bone at the Wrist
Date
2013-09-17
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
ORCID
Type
Degree Level
Masters
Abstract
The purpose of this study is to determine the correlation between texture features of Di raction
Enhanced Imaging (DEI) images and trabecular properties of human wrist bone in the assessment
of osteoporosis. Osteoporosis is a metabolic bone disorder that is characterized by reduced bone
mass and a deterioration of bone structure which results in an increased fracture risk. Since the
disease is preventable, diagnostic techniques are of major importance. Bone micro-architecture and
Bone mineral density (BMD) are two main factors related to osteoporotic fractures. Trabecular
properties like bone volume (BV), trabecular number (Tb.N), trabecular thickness (Tb.Th), bone
surface (BS), and other properties of bone, characterizes the bone architecture. Currently, however,
BMD is the only measurement carried out to assess osteoporosis. Researchers suggest that bone
micro-architecture and texture analysis of bone images along with BMD can provide more accuracy
in the assessment.
We have applied texture analysis on DEI images and extracted texture features. In our study,
we used fractal analysis, gray level co-occurrence matrix (GLCM), texture feature coding method
(TFCM), and local binary patterns (LBP) as texture analysis methods to extract texture features.
3D Micro-CT trabecular properties were extracted using SkyScanTM CTAN software. Then, we
determined the correlation between texture features and trabecular properties. GLCM energy fea-
ture of DEI images explained more than 39% of variance in bone surface by volume ratio (BS/BV),
38% of variance in percent bone volume (BV/TV), and 37% of variance in trabecular number
(Tb.N). TFCM homogeneity feature of DEI images explained more than 42% of variance in bone
surface (BS) parameter. LBP operator - LBP 11 of DEI images explained more than 34% of vari-
ance in bone surface (BS) and 30% of variance in bone surface density (BS/TV). Fractal dimension
parameter of DEI images explained more than 47% of variance in bone surface (BS) and 32% of
variance in bone volume (BV). This study will facilitate in the quanti cation of osteoporosis beyond
conventional BMD.
Description
Keywords
Texture, Osteoporosis, TFCM,GLCM, LBP, Fractal, Bone micro architecture, DEI
Citation
Degree
Master of Science (M.Sc.)
Department
Computer Science
Program
Computer Science