Comparison of The Performance of Denoising Filters in Improving The Quality of Low-Dose Brain PET Images With Two Dose Levels of 5% and 10%

Positron Emission tomography (PET) as a nuclear medicine imaging modality is increasingly used in clinical practice to diagnose disease and cancer. To achieve a high quality PET image for diagnostic purposes, a standard dose of radioactive tracer must be injected into the patient's body, which increases the risk of radiation damage. However, reducing the tracer dose leads to increase in noise and decrease in the signal-to-noise ratio and the quality of the PET. This paper aims to evaluate the performance of Gaussian filter, Bilateral filter, Non-Local Mean and Wavelet Transform filters in improving the quality of low-dose PET images. First, these filters were optimized separately for two dose levels, and then the optimal filters were applied to 40 PET brain images with a 5% and 10% dose levels. Finally, the performance of these filters was evaluated and compared using PSNR, RMSE criteria in the whole brain and RMSE and Bias criteria in different areas of the brain (e.g, thalamus, cerebellum, putamen, etc.).