Table 5 Quantitative comparisons between...

Table 5

Quantitative comparisons between MSAN and other methods in terms of PSNR and SSIM with the proposed our 3D confocal fluorescence imaging dataset. The first and second best performances are denoted in red and blue, respectively.

Methods	Noisy	NLM	BM3D	KSVD	EPLL	WNNM	PURE-LET	DnCNN
PSNR	37.79	45.92	47.13	44.56	44.82	44.50	38.69	46.77
SSIM	0.8728	0.9803	0.9831	0.9740	0.9742	0.9715	0.8874	0.9822
Methods	IRCNN	MemNet	Noise2Noise	MWCNN	WF-UNet	RIDNet	DPDN	MSAN
PSNR	46.61	44.97	47.13	45.44	47.09	46.14	45.74	47.36
SSIM	0.9788	0.9688	0.9821	0.9733	0.9836	0.9772	0.9741	0.9850

Methods	Noisy	NLM	BM3D	KSVD	EPLL	WNNM	PURE-LET	DnCNN
PSNR	37.79	45.92	47.13	44.56	44.82	44.50	38.69	46.77
SSIM	0.8728	0.9803	0.9831	0.9740	0.9742	0.9715	0.8874	0.9822
Methods	IRCNN	MemNet	Noise2Noise	MWCNN	WF-UNet	RIDNet	DPDN	MSAN
PSNR	46.61	44.97	47.13	45.44	47.09	46.14	45.74	47.36
SSIM	0.9788	0.9688	0.9821	0.9733	0.9836	0.9772	0.9741	0.9850

[ViewLarge]