A Method of CT Image Denoising Based on Residual Encoder-Decoder Network

Table 3

PSNR and SSIM in experiments.

Number	Index	LDCT image	WGAN [23]	RED-CNN [28]	Proposed algorithm
a	PSNR/dB	25.375	28.346	28.783	30.245
	SSIM	0.768	0.912	0.921	0.942
b	PSNR/dB	24.653	26.398	29.012	29.374
	SSIM	0.731	0.892	0.901	0.911
c	PSNR/dB	24.987	29.321	30.876	31.238
	SSIM	0.763	0.924	0.932	0.953
d	PSNR/dB	23.826	27.873	28.987	30.872
	SSIM	0.711	0.865	0.912	0.934
e	PSNR/dB	30.145	32.146	31.273	31.698
	SSIM	0.863	0.962	0.920	0.925
f	PSNR/dB	27.836	30.124	30.023	30.836
	SSIM	0.792	0.845	0.823	0.912
	PSNR/dB	26.834	29.834	29.867	31.345
	SSIM	0.723	0.835	0.844	0.902
h	PSNR/dB	26.214	30.013	29.839	31.314
	SSIM	0.719	0.862	0.843	0.909
i	PSNR/dB	22.245	26.215	27.831	30.134
	SSIM	0.712	0.821	0.873	0.901
Average	PSNR/dB	25.791	28.919	29.610	30.784
	SSIM	0.75	0.876	0.885	0.916