Medical-Image Super-Resolution Reconstruction Method Based on Residual Channel Attention Network

Liu Kewen; Ma Yuan; Xiong Hongxia; Yan Zejun; Zhou Zhijun; Liu Chaoyang; Fang Panpan; Li Xiaojun; Chen Yalei

doi:10.3788/LOP57.021014

Laser & Optoelectronics Progress, Volume. 57, Issue 2, 21014(2020)

Medical-Image Super-Resolution Reconstruction Method Based on Residual Channel Attention Network

Liu Kewen^1,2, Ma Yuan^1,2, Xiong Hongxia^3、*, Yan Zejun⁴, Zhou Zhijun⁵, Liu Chaoyang⁶, Fang Panpan^1,2, Li Xiaojun^1,2, and Chen Yalei^1,2

Author Affiliations

¹School of Information Engineering, Wuhan University of Technology, Wuhan, Hubei 430070, China

²Hubei Key Laboratory of Broadband Wireless Communication and Sensor Networks, Wuhan University of Technology, Wuhan, Hubei 430070, China

³School of Civil Engineering & Architecture, Wuhan University of Technology, Wuhan, Hubei 430070, China

⁴Department of Urology, Ningbo First Hospital, Key Laboratory of Translational Medicine of Urological Diseases in Ningbo, Ningbo, Zhejiang 315010, China

⁵Department of Urology, the First People''s Hospital of Tianmen, Tianmen, Hubei 431700, China

⁶State Key Laboratory of Magnetic Resonance and Atomic Molecular Physics, Wuhan Institute of Physics and Mathematics, Chinese Academy of Sciences, Wuhan, Hubei 430071, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(8)

Fig. 1. Basic unit of residual learning

Download full size

Fig. 2. Channel-attention mechanism block

Download full size

Fig. 3. Network structure based on deep residual channel attention. (a) Basic unit; (b) network structure

Download full size

Fig. 4. Loss curve of each method on lung CT dataset

Download full size

Fig. 5. Loss curve of each method on prostate MRI dataset

Download full size

Fig. 6. Comparison of rendering of images with super-resolution magnification of 2 under each super resolution method. (a) Lung tip tra CT; (b) lung leaf tra CT; (c) prostateX-0061 T2_tse_tra MRI; (d) prostateX-0082 T2_tse_tra MRI

Download full size

Table 1. Experimental environment parameters
View table
Table 1. Experimental environment parameters
Hardwareconfiguration Parameter
CPURAMGPUGPU MemoryDevelopment Framework Intel Xeon E3-1231V316G1070Ti8GPytorch1.1

Table 2. Objective evaluation of each super-resolution method

View table

Table 2. Objective evaluation of each super-resolution method

Method	Lung CT image testing set					Prostate MRI image testing set
Method	MSE	SNR /dB	PSNR /dB	SSIM	Time /s	MSE	SNR /dB	PSNR /dB	SSIM	Time /s
Bilinear	366.0564	3.8746	22.7222	0.72824	0.0498	233.7763	6.0589	24.457	0.79361	0.0380
Bicubic	286.6879	4.3645	23.8058	0.79107	0.0505	180.016	7.1662	25.5926	0.85839	0.0406
ESPCN	148.2858	4.9720	27.1032	0.85065	0.1774	84.8268	7.6330	28.8576	0.90677	0.1542
SRCNN	138.0596	5.1149	27.5442	0.85281	0.2471	73.1347	7.9603	29.5002	0.90909	0.2146
FSRCNN	136.5873	5.1170	27.6013	0.85370	0.2579	70.3323	8.1547	29.6716	0.91008	0.2226
SRResNet	100.8272	5.7331	29.6335	0.86773	0.3402	20.5025	9.7195	35.0589	0.93164	0.3495
VDSR	100.7198	5.9098	29.7222	0.86756	0.3698	19.9907	10.1670	35.1910	0.92892	0.3794
EDSR	100.6880	6.0319	29.8013	0.87068	0.3944	19.0991	10.3805	35.3742	0.93266	0.3828
Proposed	94.8284	6.0913	29.9796	0.87213	0.4321	17.1458	10.7339	35.8514	0.9345	0.4226

Tools

Get Citation

Copy Citation Text

Liu Kewen, Ma Yuan, Xiong Hongxia, Yan Zejun, Zhou Zhijun, Liu Chaoyang, Fang Panpan, Li Xiaojun, Chen Yalei. Medical-Image Super-Resolution Reconstruction Method Based on Residual Channel Attention Network[J]. Laser & Optoelectronics Progress, 2020, 57(2): 21014

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites