Related Experiment Videos
[Characterization of the difference between filtered back projection and ordered subsets expectation maximization in
A Pérez1, R Piotrkowski, R Galli
1Escuela de Ciencia y Tecnología de la Universidad Nacional de General San Martín, Alem 3901, Villa Ballester, Argentina.
Revista Espanola De Medicina Nuclear
|January 14, 2004
Summary
The Ordered Subsets Expectation Maximization (OSEM) algorithm enhances nuclear medicine image quality by reducing Poisson noise and annular artifacts compared to Filtered Back Projection (BP). Wavelet Transform effectively analyzes these improvements.
Area of Science:
- Medical Imaging
- Signal Processing
- Nuclear Medicine
Context:
- Reconstruction algorithms are crucial for high-quality nuclear medicine tomographic images.
- Filtered Back Projection (BP) is an analytical method, while Ordered Subsets Expectation Maximization (OSEM) is a statistical method.
- Image perturbations like Poisson noise, annular artifacts, and attenuation affect image fidelity.
Purpose:
- To compare the efficiency of OSEM and BP reconstruction algorithms using Wavelet Transform (WT).
- To analyze the impact of WT on identifying and quantifying image artifacts and noise.
- To evaluate the performance of OSEM versus BP in mitigating common image perturbations.
Summary:
- Wavelet Transform (WT) on the Haar basis was used to compare BP and OSEM algorithms on homogeneous nuclear medicine images.
- OSEM demonstrated superior performance in filtering Poisson noise and reducing annular artifacts, particularly near the rotation center.
- Attenuation effects were similar, with OSEM showing slightly higher peripheral activity than BP.
Impact:
- OSEM significantly improves image quality in nuclear medicine by reducing noise and artifacts.
- Wavelet Transform proves to be a valuable tool for artifact identification and analysis in tomographic imaging.
- Further research using WT can help mitigate the deleterious effects of artifacts on nuclear medicine images.