240
Bulletin UASMV, Veterinary Medicine, 69(1-2)/2012
Print ISSN 1843-5270; Electronic ISSN 1843-5378
Using Fractal Analysis to Assess the Ultrasonographic Images
Andrei ȘTEFAN, Mircea MIRCEAN
University of Agricultural Sciences and Veterinary Medicine, Faculty of Veterinary Medicine Mănăştur
Street, No. 3-5, 400372, Cluj-Napoca, Romania
stefan.andrei.vet@gmail.com
Abstract: Ultrasonography is used in Veterinary Medicine to evaluate the structure of internal organs.
Interpretation of ultrasound images is dependent on the evaluators experience, the machine settings or the
frequency at which the scan is made. In order to assess more objectively the interpretation of digital
ultrasound images, we propose the use of Fractal Analysis (FA) as a complementary method.
Keywords: fractal dimension, fractal analysis, ultrasonography, liver structure.
Introduction Ultrasonography is used in Veterinary Medicine to evaluate the structure
of internal organs. Interpretation of ultrasound images is dependent on the evaluators experience,
the machine settings or the frequency at which the scan is made. Recent research papers show
that the FA might by an efficient tool in characterizing digital medical images regarding their
texture (Chen D.-R., 2005). The result of FA is represented by the Fractal Dimension (FD) which
is a numerical value between 1 and 2 (Irini S. Reljin and Reljin B. D., 2002).
Aims and objectives. To provide an objective evaluation of ultrasound images, we
proposed the FA of digital ultrasound images, along with standard assessment protocol.
Materials and methods. Digital ultrasound images were obtained from one healthy dog
and from three dogs diagnosed with different liver pathologies (metastases from the breast,
hemangiosarcoma and lymphoma) (Tab 1) in the Clinic of Internal Medicine, Faculty of
Veterinary Medicine, U.S.A.M.V. Cluj-Napoca, Romania.
Tab 1
Number of digital ultrasound images taken for each case
Normal liver Liver metastasis Hemangiosarcoma Lymphoma
No. of digital ultrasound
images taken
6 4 16 6
No. of ROIs 16 16 16 16
In order to achieve Fractal Analysis an equal number of square shaped regions of
interest (ROIs) were cropped using specialized software PhotoScape
©
(MOOII TECH
©
). The
ROIs were then loaded in the ImageJ™ (Wayne Rasband (HIH), U.S.A.) software and processed
by the box-counting method.
The resulted data were recorded and statistically analyzed using the M.S. Excel™
(Microsoft™, U.S.A.) and GraphPad InStat™ v3.05 (GraphPad™, U.S.A.) software, by
comparing the pathological data against the normal data.
brought to you by CORE View metadata, citation and similar papers at core.ac.uk