Industrial x-ray image enhancement network based on a ray scattering model

Xiaxu Ding; Xiaxu Ding; Yi Liu; Yi Liu; Hongxu Yan; Hongxu Yan; Pengcheng Zhang; Pengcheng Zhang; Niu Guo; Niu Guo; Zhiguo Gui; Zhiguo Gui

doi:10.1364/AO.493750

Applied Optics
Vol. 62,
Issue 20,
pp. 5526-5537
(2023)
•https://doi.org/10.1364/AO.493750

Industrial x-ray image enhancement network based on a ray scattering model

Xiaxu Ding, Yi Liu, Hongxu Yan, Pengcheng Zhang, Niu Guo, and Zhiguo Gui

Not Accessible

Your library or personal account may give you access

Get PDF
Email
Share
Get Citation
Copy Citation Text
Xiaxu Ding, Yi Liu, Hongxu Yan, Pengcheng Zhang, Niu Guo, and Zhiguo Gui, "Industrial x-ray image enhancement network based on a ray scattering model," Appl. Opt. 62, 5526-5537 (2023)

Export Citation
- BibTex
- Endnote (RIS)
- HTML
- Plain Text
Citation alert
Save article

Check for updates

Abstract

X-ray images frequently have low contrast and lost edge features because of the complexity of objects, attenuation of reflected light, and scattering superposition of rays. Image features are frequently lost in traditional enhancement methods. In this paper, we use a ray scattering model to estimate coarsely clear images and an encoder–decoder network and multi-scale feature extraction module to add multi-scale and detail information to the images. To selectively emphasize useful features, a dual attention module and UnsharpMasking with learnable correction factors are used. The results of the experiments demonstrate that the method may significantly enhance the quality of x-ray images.

Full Article | PDF Article

More Like This

EDIRNet: an unsupervised deformable registration model for X-ray and neutron images

Qingtian Zeng, Congli Yang, Quan Gan, Qihong Wang, and Shansong Wang
Appl. Opt. 62(29) 7611-7620 (2023)

Underwater motion scene image restoration based on an improved U-Net network

Jianye Liu, Hu Luo, and Dawei Tu
Appl. Opt. 63(1) 228-238 (2024)

Image dehazing algorithm based on optimized dark channel and haze-line priors of adaptive sky segmentation

Guangmang Cui, Qiong Ma, Jufeng Zhao, Shunjie Yang, and Ziyi Chen
J. Opt. Soc. Am. A 40(6) 1165-1182 (2023)

Previous Article Next Article

Data availability

Data underlying the results presented in this paper are available in Ref. [35].

35. B. Li, W. Ren, D. Fu, D. Tao, D. Feng, W. Zeng, and Z. Wang, “Benchmarking single-image dehazing and beyond,” IEEE Trans. Image Process. 28, 492–505 (2018). [CrossRef]

Cited By

You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Figures (10)

You do not have subscription access to this journal. Figure files are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Tables (5)

You do not have subscription access to this journal. Article tables are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Equations (13)

You do not have subscription access to this journal. Equations are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Conv Layer	1	2	3	4	5
Kernel size	$1 \times 1$	$3 \times 3$	$5 \times 5$	$7 \times 7$	$3 \times 3$
Receptive field	1	3	7	13	15
Cout	3	3	3	3	3

Encoder					Decoder
Layer	Kernel	Cin	Cout	Nums	Layer	Kernel	Cin	Cout	Nums
Conv	$3 \times 3$	1	8	1	Conv	$1 \times 1$	256	128	1
Conv	$3 \times 3$	8	8	2	Upsample	$2 \times 2$	-	-	1
Maxpool	$2 \times 2$	-	-	1	Conv	$3 \times 3$	128	64	1
Conv	$3 \times 3$	8	16	1	Conv	$3 \times 3$	64	64	2
Conv	$3 \times 3$	16	16	2	Upsample	$2 \times 2$	-	-	1
Maxpool	$2 \times 2$	-	-	1	Conv	$3 \times 3$	64	32	2
Conv	$3 \times 3$	16	32	1	Conv	$3 \times 3$	32	32	2
Conv	$3 \times 3$	32	32	2	Upsample	$2 \times 2$	-	-	1
Maxpool	$2 \times 2$	-	-	1	Conv	$3 \times 3$	32	16	2
Conv	$3 \times 3$	32	64	1	Conv	$3 \times 3$	16	16	2
Conv	$3 \times 3$	64	64	2	Upsample	$2 \times 2$	-	-	1
Maxpool	$2 \times 2$	-	-	1	Conv	$3 \times 3$	16	8	2
Conv	$3 \times 3$	64	128	1	Conv	$3 \times 3$	8	8	2
Conv	$3 \times 3$	128	128	2	Conv	$1 \times 1$	8	1	1

Method	Indices	Input	CLAHE [10]	DCP [13]	END-GTM [20]	AOD [21]	MMTMO [36]	Ours
Exp1	EOG	3.48	2.06	2.78	5.69	5.46	7.25	9.86
	SD	59.29	68.28	67.10	72.77	73.63	77.17	80.68
	Entropy	4.56	5.05	5.06	4.35	4.02	4.46	4.58
	NRSS	0.69	0.80	0.83	0.69	0.72	0.85	0.88
Exp2	EOG	3.56	3.86	4.78	6.24	7.89	9.25	11.92
	SD	72.02	68.35	64.54	70.61	71.63	77.24	79.68
	Entropy	4.79	5.27	5.04	5.18	5.00	5.23	5.49
	NRSS	0.68	0.77	0.71	0.68	0.69	0.83	0.89
Exp3	EOG	2.62	2.36	3.91	8.73	9.54	10.25	14.82
	SD	70.37	72.93	72.36	78.40	79.98	83.15	87.42
	Entropy	4.24	5.70	4.31	5.81	4.61	5.82	6.35
	NRSS	0.65	0.71	0.72	0.69	0.75	0.85	0.92
Exp4	EOG	2.28	2.54	3.85	8.69	9.24	10.82	13.76
	SD	72.68	74.39	75.73	78.23	80.49	83.27	89.63
	Entropy	4.62	5.08	5.65	6.14	6.87	6.49	7.25
	NRSS	0.67	0.73	0.72	0.70	0.73	0.83	0.90

	Parameter	Time
CHAHE	/	0.0049 (s)
DCP	/	0.0920 (s)
AODNet	1537	0.0014 (s)
EDN-GTM	522823	0.0040 (s)
MMTMO	/	0.0551 (s)
Ours	587073	0.0062 (s)

	EOG	SD	Entropy	NRSS
W/o Encoder–decoder network and multi-scale feature extraction module	7.13	70.92	3.76	0.76
W/o Spatial and channel dual attention module	9.06	73.24	4.13	0.80
W/o UM	9.32	78.54	4.22	0.85
Proposed method	9.86	80.68	4.58	0.88

Abstract

Data availability

Cited By

Figures (10)

Tables (5)

Equations (13)

Applied Optics