Parallel multifeature extracting network for infrared image enhancement

Table 1. Parameters of SFW map blocks
View table
View in Article
Table 1. Parameters of SFW map blocks
Type K s p c
Conv 1 1 0 32
conv 3 1 1 32
Conv 5 1 2 32
Conv 3 1 1 32

Table 2. Parameters of TSFEB
View table
View in Article
Table 2. Parameters of TSFEB
Path Type K s p c
2_1 Conv 3 2 1 48
2_2 Conv 1 1 0 32
2_3 MS-FE / / / 62
2_4 MS-FE / / / 96
2_5 Conv 3 1 1 96
3_1 Conv 4 4 0 96
3_2 Conv 1 1 0 64
3_3 AB / / / 64
3_4 Conv 3 1 2 64
3_5 Conv 3 1 2 64
3_6 Conv 3 1 2 64
3_7 AB / / / 64
3_8 deconv 4 2 1 96

Table 3. Test result on BSD200 with \begin{document}${\boldsymbol{\alpha}} $\end{document}∈ [0.5, 0.51]
View table
View in Article
Table 3. Test result on BSD200 with \begin{document}${\boldsymbol{\alpha}} $\end{document}∈ [0.5, 0.51]
Method HE CLAHE SSR MSR TEN TIECNN IE-GAN Proposed
PSNR 15.95 22.19 16.51 17.57 25.07 24.60 26.23 35.42
SSIM 0.72 0.93 0.88 0.90 0.82 0.80 0.92 0.95

Table 4. Test result on real infrared images with \begin{document}${\boldsymbol{\alpha}}$\end{document}∈ [0.5, 0.51]
View table
View in Article
Table 4. Test result on real infrared images with \begin{document}${\boldsymbol{\alpha}}$\end{document}∈ [0.5, 0.51]
Method HE CLAHE SSR MSR TEN TIECNN IE-GAN Proposed
PSNR 13.06 24.78 15.89 18.36 25.77 23.25 26.85 35.72
SSIM 0.53 0.95 0.89 0.93 0.89 0.88 0.94 0.96

Table 5. PSNR and SSIM on BSD200 with different \begin{document}${\boldsymbol{\alpha}}$\end{document} using proposed method

View table

View in Article

Table 5. PSNR and SSIM on BSD200 with different \begin{document}${\boldsymbol{\alpha}}$\end{document} using proposed method

$ \alpha $	$ \left[\mathrm{0.1,0.11}\right] $	$ \left[\mathrm{0.2,0.21}\right] $	$ \left[\mathrm{0.3,0.31}\right] $	$ \left[\mathrm{0.4,0.41}\right] $
PSNR	30.5762	31.0370	28.1290	32.0594
SSIM	0.8343	0.8867	0.9074	0.9058

Table 6. Ablation experiments

View table

View in Article

Table 6. Ablation experiments

Method	PSNR	SSIM	Time/s
Path1(SFW)	27.35	0.92	0.05
Path2 without MS-feature extraction	25.26	0.83	0.09
Path2	27.03	0.87	0.14
Path3	24.58	0.92	0.14
TSFEB	29.15	0.90	0.21
SFW + TSFEB without MS-feature extraction	30.99	0.93	0.22
SFW + TSFEB	35.42	0.95	0.26

Tools

Get Citation

Copy Citation Text

Zhongxiang Pang, Xie Liu, Guihua Liu, Yinjun Gong, Han Zhou, Hongwei Luo. Parallel multifeature extracting network for infrared image enhancement[J]. Infrared and Laser Engineering, 2022, 51(8): 20210957

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Infrared technology and application

Received: Dec. 13, 2021

Accepted: --

Published Online: Jan. 9, 2023

The Author Email:

DOI:10.3788/IRLA20210957

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Parameters of SFW map blocks

Table 1. Parameters of SFW map blocks

Table 2. Parameters of TSFEB

Table 2. Parameters of TSFEB

Table 3. Test result on BSD200 with \begin{document}${\boldsymbol{\alpha}} $\end{document}∈ [0.5, 0.51]

Table 3. Test result on BSD200 with \begin{document}${\boldsymbol{\alpha}} $\end{document}∈ [0.5, 0.51]

Table 4. Test result on real infrared images with \begin{document}${\boldsymbol{\alpha}}$\end{document}∈ [0.5, 0.51]

Table 4. Test result on real infrared images with \begin{document}${\boldsymbol{\alpha}}$\end{document}∈ [0.5, 0.51]

Table 5. PSNR and SSIM on BSD200 with different \begin{document}${\boldsymbol{\alpha}}$\end{document} using proposed method

Table 5. PSNR and SSIM on BSD200 with different \begin{document}${\boldsymbol{\alpha}}$\end{document} using proposed method

Table 6. Ablation experiments

Table 6. Ablation experiments