E1-2 - Sedma Nacionalna Konferencija so Me|unarodno U~estvo...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
A WAVELET APPROACH ON NM IMAGES FILTERING USING ADJACENT IMAGES INFORMATION Cvetko D. Mitrovski 1 , Mitko B. Kostov 1 1 University St. Kliment Ohridski, Faculty of Technical Sciences, Bitola – Macedonia, [email protected] , [email protected] Abstract – In this paper we present our approach on pre-processing chest region dynamic NM images. It enables anatomical data extraction of the vena cava and the heart. The aim of the method is developing sophisticated diagnostic software that could automatically offer the optimal positions and the shapes of the regions of interest needed for heart studies. Key words – Nuclear medicine image, wavelet- domain filtering, autocorrelation, denoising. 1. INTRODUCTION Nuclear Medicine (NM) images are diagnostic digital images, which provide both anatomical and functional information. They present the projection of the distribution of radioisotope(s) in a body of a patient after injection of adequate dose of radioisotope(s). The raw NM images are created by accumulating the emitted gamma rays from a patient over a fixed observation period by computerized gamma cameras. They have a low signal-to-noise ratio (SNR) due to the nature of the gamma ray emission process and the operational characteristics of the gamma cameras (low count levels, scatter, attenuation, and electronic noises in the detector/camera). The noise obeys a Poisson law and is highly dependent on the space distribution of the image signal intensity. Therefore, a suitable image pre-processing must precede the NM images analysis in order to provide an accurate recognition of the anatomical data of the patient (the boundaries of the various objects – organs). This process of separating signal from noise is a rather difficult and much diversified task that should be adjusted to the organs and tissues, which physiology is to be investigated. In [1], [2], [3], [4] and [5] we proposed several approaches to cope with this problem. In [1] the whole process of spreading of the radionuclide is divided in three successive phases and the images that belong to one specific phase are processed separately from the others. The processing includes changing images resolution and applying autocorrelation technique. In [2] the images are filtered by applying the wavelet shrinkage program, where the set threshold is same for all the wavelet coefficients in one level. In [3] the images’ denoising is carried out by modifying images histogram. In [4] we try to denoise images by filtering in the direction that is normal to the radionuclide spreading direction. In [5] we combine DWT realized via QMF filters with a specific strategy for selecting an appropriate threshold. Due to the signal-dependence of the Poisson noise, the Anscombe variance-stabilizing transformation is applied.
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 02/18/2010 for the course ITK ETF113L07 taught by Professor Popovskiborislav during the Spring '10 term at Pacific.

Page1 / 6

E1-2 - Sedma Nacionalna Konferencija so Me|unarodno U~estvo...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online