Hyperspectral Image Reconstruction Based on Improved Residual Dense Network

Yong Li; Qiuyu Jin; Huaici Zhao; Bo Li

doi:10.3788/AOS202141.0730001

Acta Optica Sinica, Volume. 41, Issue 7, 0730001(2021)

Hyperspectral Image Reconstruction Based on Improved Residual Dense Network

Yong Li¹, Qiuyu Jin^1,2, Huaici Zhao^2、*, and Bo Li³

Author Affiliations

¹School of Electrical Engineering, Shenyang University of Technology, Shenyang, Liaoning 110870, China

²Key Laboratory of Optical-Electronics Information Processing, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, Liaoning 110016, China

³College of Information, Shenyang Institute of Engineering, Shenyang, Liaoning 110136, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(8)

Fig. 1. Schematic of reconstruction process

Download full size

Fig. 2. Schematic of RDSB

Download full size

Fig. 3. Schematic of network structure

Download full size

Fig. 4. Reconstruction results of some images in ICVL dataset. (a)Algorithm in Ref. [11]; (b) algorithm in Ref. [12]; (c) algorithm in Ref. [17]; (d) algorithm in Ref. [18]; (e) proposed algorithm; (f) original RGB images

Download full size

Fig. 5. Reconstruction results of different algorithms in CAVE dataset. (a) Algorithm in Ref. [11]; (b) algorithm in Ref. [12]; (c) algorithm in Ref. [17]; (d) algorithm in Ref. [18]; (e) proposed algorithm; (f) original images

Download full size

Fig. 6. Reconstruction images on real-world images taken by an ordinal camera. (a) Algorithm in Ref. [11]; (b) algorithm in Ref. [12]; (c) algorithm in Ref. [17]; (d) algorithm in Ref. [18]; (e) proposed algorithm; (f) original RGB image

Download full size

Table 1. Evaluate results of ICVL dataset
View table
Table 1. Evaluate results of ICVL dataset
Indicator Algorithm in Ref. [11] Algorithm in Ref. [12] Algorithm in Ref. [17] Algorithm in Ref. [18] Proposed algorithm
MRAE 0.0614 0.0435 0.0419 0.0331 0.0327
RMSE 2.632 1.956 1.778 1.506 1.452

Table 2. Evaluate results of CAVE dataset

View table

Table 2. Evaluate results of CAVE dataset

Dataset	Algorithm in Ref.[11]		Algorithm in Ref.[12]		Algorithm in Ref.[17]		Algorithm in Ref.[18]		Proposed algorithm
Dataset	MRAE	RMSE	MRAE	RMSE	MRAE	RMSE	MRAE	RMSE	MRAE	RMSE
Paints subset	0.0887	5.372	0.0755	4.532	0.0742	4.313	0.0667	4.305	0.0641	4.295
Food subset	0.0675	3.297	0.0524	3.125	0.0489	2.818	0.0365	2.607	0.0383	2.536
Real subset	0.0617	3.120	0.0631	3.327	0.0601	3.547	0.0579	3.004	0.0582	2.974
Skin subset	0.0804	3.781	0.0656	3.498	0.0646	3.414	0.0615	3.262	0.0612	3.245
Stuff subset	0.0973	4.668	0.0706	3.964	0.0709	4.236	0.0652	3.682	0.0641	3.503

Tools

Get Citation

Copy Citation Text

Yong Li, Qiuyu Jin, Huaici Zhao, Bo Li. Hyperspectral Image Reconstruction Based on Improved Residual Dense Network[J]. Acta Optica Sinica, 2021, 41(7): 0730001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites