Denoising of Medical Images Using Image Fusion Techniques

Abstract

Medical Imaging is playing the key role in diagnosing and treatment of diseases such as locating the tumours in brain, thin fractures in bones, detection of cancer cells in early stages etc. For making accurate decisions the images acquired by various medical imaging modalities must be free from noise. So Image denoising became an important pre-processing step in Medical image analysis. Developing the denoising algorithms is a difficult task because diagnostic information must be preserved while removing the noise. Earlier the denoising algorithms were designed in the spatial domain such as median filtering, harmonic filtering, and weiner filtering etc. by directly working on the pixel values, these methods will remove noise while introducing the blur in the denoised images. At present advanced mathematical models such as Partial differential equations which are useful for edge preservation and multiresolution analysis useful for preserving directional oriented information and texture became very popular in developing the denoising techniques. In this paper we are using Total Variational approach (PDE method) and Complex Dual Tree wavelet transform (Multiresolution analysis)method to denoise the medical images and we perform the fusion of the two denoised images resulting from the above denoising techniques. The performance of the proposed algorithm is compared with the existing methods using PSNR (Peak Signal to Noise Ratio) and SSIM (Structural Similarity Index.). The performance of the fusion process is measured using MI (Mutual Information), Edge Association and Spatial Frequency (SF) measures. The results showed that the proposed method is having better PSNR values diagnostically acceptable and very much useful for the diagnosis& treatment phases.