Denoising Method for MRI Images Using Modified BM3D Filter with Complex Network and Artificial Neural Networks

Table 1

Summary of studies on denoise MRI image.

Author	Denoise method	Noise	Dataset	Metrics highest
Chang et al. 2019 [8]	Bilateral filter & neural network	Gauss	Local datasets	Gauss 1%	PSNR: 39.29
					SSIM: 0.983
Tripathi et al. 2020 [9]	CNN	Rician	Local datasets and brainweb dataset	Rician 1%	PSNR: 43.18
					SSIM: 0.987
Moreno López et al. 2021 [10]	Unsupervised learning	Standard deviation of the noise	Local datasets: 1172 MRI images and brainweb dataset	Standard deviation of the noise σ = 50	PSNR: 38.015
					SSIM: 0.8977
Sreelakshmi et al. 2021 [11]	Adaptive median filter and CNN	Gauss, sat and pepper, shrinking	Local datasets	Gauss 10%	PSNR: 55.59
				Gauss 50%	PSNR: 48.68
				Shrinking 10%	PSNR: 68.85
				SSIM: 0.989
Wang et al. 2022 [12]	Nonlocal structural similarity and low-rank sparse representation	Rician	Brainweb 3D T1-weighted	Rician 4%	PSNR: 38.503
					SSIM: 0.976
Mehta et al. 2022 [13]	U-NET architecture	Gauss	253 MRI images	Gauss 25%	PSNR: 30.96
Kollem et al. 2023 [14]	Diffusivity function	Noise free image + poisson noise	BraTS2020	PSNR: 42.78
				SSIM: 0.99645