A Network with Composite Loss and Parameter-free Chunking Fusion Block for Super-Resolution MR Image

Table 2

The super-resolution reconstructed LR image is obtained by k-space truncation (TD). The experimental data show the comparison of other methods with our proposed method in various evaluation indexes: PSNR, SSIM, and RMSE.

	k-space truncation (TD)
	PD			T1			T2
	PSNR	SSIM	RMSE	PSNR	SSIM	RMSE	PSNR	SSIM	RMSE
Bicubic	32.34	0.9555	0.0252	31.65	0.9361	0.0270	32.36	0.9477	0.0246
ESPCNN [42]	31.73	0.9525	0.0270	31.02	0.9273	0.0287	31.71	0.9427	0.0263
VRCNN [48]	32.86	0.9621	0.0238	32.10	0.9452	0.0257	33.67	0.9613	0.0213
EDSR [25]	34.14	0.9736	0.0209	32.70	0.9529	0.0240	35.49	0.9724	0.0174
PAN [49]	33.86	0.9726	0.0216	32.75	0.9580	0.0240	35.05	0.9703	0.0183
NAFNet [50]	32.32	0.9574	0.0252	31.67	0.9395	0.0269	32.42	0.9523	0.0244
Proposed	34.20	0.9752	0.0208	32.87	0.9591	0.0238	35.66	0.9734	0.0172

Bold values highlight the optimal results.