Loading...

Improvement of PET Images Using Wavelet Method and its Comparison with Usual Clinical Method

Mohammadi, Mohammad Sadegh | 2020

442 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 53054 (46)
  4. University: Sharif University of Technology
  5. Department: Energy Engineering
  6. Advisor(s): Hosseini, Abolfazl
  7. Abstract:
  8. Positron emission tomography (PET) imaging is a cross-sectional imaging method in nuclear medicine. In recent years, there have been increasing interest in this method.PET imaging is very sensitive and the presence of noise in the results decreases the the image quality and precision of detection. Thus it is critical to to investigate noise in PET systems with noisy images, since it could decrease the system’s ability to detect tumors. The common method to decrease noise in nuclear medicine images is the Fourier transform (FT). The method has many applications in image processing and analysis. However, in the FT method, information regarding time and position are lost in the frequency domain. On the other hand, in the FT method, the signal needs to be constant, while signals in the nuclear medicine are not constant. It is therefor important to find an alternative method without these shortcomings. One of the option is the wavelet transform. In this study we aim to find the effect of wavelet to decrease noise in PET systems’ images. For this purpose, the best wavelet transform method was found by implementing various wavelet transform methods on PET images, and the method was implemented for four body parts brain, thoracic cage, abdominal regions, and pelvic regions. The required programs for image reconstruction and analysis and noise reduction were developed using MATLAB. The effect of 75 wavelets of the families Daubechies, Coiflets, and Symlets were investigated on images using four different methods. To compare the denoised images with the reference image, the main indices SSIM and Line profile were used.Based on the SSIM values, the best denoising method was found to be the SWT by the surface-dependent thresholding method in the first analysis layer. In this method, SSIM reaches its maximum and the similarity was highest between the denoised and the reference images
  9. Keywords:
  10. Nuclear Medicine ; Noise Reduction ; Fourier Transform ; Wavelet Transform ; Positron Emission Tomography (PET)

 Digital Object List

 Bookmark

No TOC