A Combined Algorithm for Noisy Whole-body Bone Scan Image Enhancement

Abstract

A bone scan is one of the most important diagnoses done using a gamma camera in nuclear medicine for detecting bone problems, such as cancer lesion, to find out if the cancers spread to the bones besides finding out how the metastases respond to chemotherapy and radiotherapy treatment. Generally nuclear medicine images degraded by a large amount of noise, which is effecting on the resulting image. In this work we work we want to propose a combined enhancement algorithm based on fast Fourier transform and Sobel, for metastases whole-body bone scan image enhancement, to reduce the image noise, increase the image quality for better viewing and assist the nuclear medicine physician diagnosing images effectively. The proposed algorithm compared with existing enhancement algorithms such as histogram equalization, adaptive histogram equalization, log transformation and gamma correction. The algorithm applied to seven patients with bone metastases. It turns out that the proposed algorithm can help to improve the quality and visualization of the images. Our simulations show that the proposed algorithm removes the noise without significant blurring the structure of the image, increase the image quality and gives clear legions. The algorithms evaluated by calculating PSNR and RMSE. The proposed algorithm gives higher PSNR and lower RMSE.