Tampere University of Technology

TUTCRIS Research Portal

Textural Features in Medical Magnetic Resonance Image Analysis of the Brain and Thigh Muscles

Research output: Book/ReportDoctoral thesisCollection of Articles

Details

Original languageEnglish
PublisherTampere University of Technology
Number of pages64
ISBN (Electronic)978-952-15-3823-0
ISBN (Print)978-952-15-3816-2
Publication statusPublished - 21 Oct 2016
Publication typeG5 Doctoral dissertation (article)

Publication series

Name Tampere University of Technology. Publication
Volume1418
ISSN (Print)1459-2045

Abstract

Magnetic resonance imaging (MRI) provides high-quality images with excellent contrast detail of soft tissues and anatomic structures. MR images contain a large amount of detailed information – some of which is invisible to the human eye. Detailed information can be analysed with computer-assisted texture analysis (TA), which is based on features describing the grey level relationships between image pixels.
The aim of this thesis was to assess the information content of textural features based on the image histogram, grey level co-occurrence matrix, and grey level run-length matrix. The strengths and limitations of the various textural features in medical MR image analysis were evaluated. The study was conducted by analysing different clinical data with TA in the clinical environment, and the results of the learning process were then gathered in this thesis.
Our results indicated that all features have limitations in terms of their discrimination capacity in medical MR images and their dependence on the size of the region of interest and MR imaging parameters. By considering these limitations, TA may help in various MR imaging applications by revealing textural information of the images of various human organs.

Field of science, Statistics Finland

Downloads statistics

No data available