Study on the Effect of Denoising Algorithms on the Parametric Maps of IVIM Imaging

Intra Voxel Incoherent Motion Magnetic Resonance Imaging (IVIM-MRI) is a quantitative imaging method used for the diagnosis of pathological disorders. The accuracy of the parameters derived from the IVIM images significantly affects the precision of diagnosis. Many researches are in progress in this area, aiming the improvement of the accuracy of IVIM parameters. The accuracy of IVIM parameters is affected by the presence of noise in IVIM images, which results in poor Signal to Noise Ratio (SNR). The noise effect becomes more significant in IVIM images where high diffusion weights are used. The proposed work is a preliminary study to analyze the effect of various denoising filters applied to the noisy parametric maps derived from IVIM images. The Gaussian smoothing filter, Non local Means filter (NLM), Anisotropic Diffusion filter (AD) and Bilateral filter are considered for comparison. The results show that, denoising the parametric maps will improve the signal intensity and it is observed that NLM filter shows better results in terms of both qualitative and quantitative metrics. Moreover, our work proved that the experimentation studies in phantom data is valid and can be used in the absence of adequate number of clinical IVIM dataset.