Continuous digital zooming using local self-similarity-based super-resolution for an asymmetric dual camera system

Byeongho Moon; Soohwan Yu; Seungyong Ko; Seonhee Park; Joonki Paik

doi:10.1364/JOSAA.34.000991

Journal of the Optical Society of America A
Vol. 34,
Issue 6,
pp. 991-1003
(2017)
•https://doi.org/10.1364/JOSAA.34.000991

Continuous digital zooming using local self-similarity-based super-resolution for an asymmetric dual camera system

Byeongho Moon, Soohwan Yu, Seungyong Ko, Seonhee Park, and Joonki Paik

Not Accessible

Your library or personal account may give you access

Get PDF
Email
Share
Get Citation
Copy Citation Text
Byeongho Moon, Soohwan Yu, Seungyong Ko, Seonhee Park, and Joonki Paik, "Continuous digital zooming using local self-similarity-based super-resolution for an asymmetric dual camera system," J. Opt. Soc. Am. A 34, 991-1003 (2017)

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

Check for updates

Abstract

This paper presents a digital zooming method using a super-resolution (SR) algorithm based on the local self-similarity between the wide- and tele-view images acquired by an asymmetric dual camera system. The proposed SR algorithm consists of four steps: (i) registration of an optically zoomed image to the wide-view image, (ii) restoration of the central region of the zoomed wide-view image, (iii) restoration of the boundary region of the zoomed wide-view image, and (iv) fusion of the results from steps (ii) and (iii). Since an asymmetric dual camera system acquires different-resolution images on the same scene due to the different optical specifications, the proposed method can restore the low-resolution wide-view image using the ideal high-frequency component estimated from the optically zoomed image. Experimental results demonstrate that the proposed method can provide significantly improved high-resolution wide-view images compared to existing single-image-based SR methods.

Full Article | PDF Article

More Like This

Single image super-resolution using locally adaptive multiple linear regression

Soohwan Yu, Wonseok Kang, Seungyong Ko, and Joonki Paik
J. Opt. Soc. Am. A 32(12) 2264-2275 (2015)

How to deal with color in super resolution reconstruction of images

Rui Gong, Yi Wang, Yilin Cai, and Xiaopeng Shao
Opt. Express 25(10) 11144-11156 (2017)

Controlled angular and radial scanning for super resolution concentric circular imaging

Xian Du and Brian Anthony
Opt. Express 24(20) 22581-22595 (2016)

Previous Article Next Article

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 (17)

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 (3)

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 (9)

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

Image	Method	[21]	[26]	[9]	[11]	[27]	[28]	Proposed
Fig. 7(a)	PSNR	26.22	27.72	28.25	29.06	28.65	29.15	30.13
Fig. 7(a)	SSIM	0.9603	0.9736	0.9753	0.9799	0.9780	0.9803	0.9834
Fig. 7(b)	PSNR	33.37	34.13	34.50	34.80	35.02	34.68	36.52
Fig. 7(b)	SSIM	0.9501	0.9596	0.9611	0.9645	0.9644	0.9640	0.9757
Fig. 7(c)	PSNR	27.80	29.06	29.04	29.47	29.52	29.79	30.02
Fig. 7(c)	SSIM	0.9531	0.9623	0.9615	0.9637	0.9641	0.9650	0.9789
Fig. 7(d)	PSNR	29.31	30.02	30.12	30.78	30.74	30.89	31.18
Fig. 7(d)	SSIM	0.9771	0.9817	0.9817	0.9841	0.9838	0.9847	0.9865
Fig. 7(e)	PSNR	28.05	29.72	30.06	30.80	30.71	31.25	31.45
Fig. 7(e)	SSIM	0.9634	0.9757	0.9746	0.9782	0.9776	0.9791	0.9825
Fig. 7(f)	PSNR	28.15	29.54	29.03	30.02	29.95	30.34	32.22
Fig. 7(f)	SSIM	0.9255	0.9453	0.9422	0.9515	0.9507	0.9526	0.9666
Fig. 7(g)	PSNR	31.38	33.11	32.87	33.78	33.91	33.79	33.94
Fig. 7(g)	SSIM	0.9819	0.9883	0.9867	0.9899	0.9901	0.9896	0.9900
Fig. 7(h)	PSNR	30.02	30.78	30.93	31.48	31.58	31.35	33.06
Fig. 7(h)	SSIM	0.9138	0.9326	0.9338	0.9418	0.9415	0.9394	0.9525

Image	Method	[21]	[26]	[9]	[11]	[27]	[28]	Proposed
Fig. 7(a)	PSNR	22.93	24.70	24.80	25.11	24.92	25.42	30.55
Fig. 7(a)	SSIM	0.9164	0.9463	0.9469	0.9504	0.9486	0.9537	0.9863
Fig. 7(b)	PSNR	29.12	30.03	30.20	30.30	30.32	30.34	37.88
Fig. 7(b)	SSIM	0.8775	0.9023	0.9054	0.9061	0.9051	0.9067	0.9874
Fig. 7(c)	PSNR	24.50	26.17	25.95	26.18	26.00	26.55	29.52
Fig. 7(c)	SSIM	0.8725	0.8950	0.8941	0.8956	0.8941	0.8993	0.9443
Fig. 7(d)	PSNR	25.48	26.29	26.44	26.57	26.54	26.70	32.87
Fig. 7(d)	SSIM	0.9405	0.9537	0.9545	0.9555	0.9550	0.9571	0.9908
Fig. 7(e)	PSNR	24.61	26.64	26.40	26.70	26.48	27.16	33.14
Fig. 7(e)	SSIM	0.9209	0.9489	0.9455	0.9496	0.9469	0.9519	0.9881
Fig. 7(f)	PSNR	23.67	25.38	24.97	25.24	25.17	25.66	31.87
Fig. 7(f)	SSIM	0.8062	0.8640	0.8587	0.8654	0.8637	0.8708	0.9738
Fig. 7(g)	PSNR	28.52	30.35	29.93	30.28	30.38	30.38	35.09
Fig. 7(g)	SSIM	0.9650	0.9772	0.9746	0.9768	0.9776	0.9766	0.9920
Fig. 7(h)	PSNR	25.70	26.41	26.56	26.72	26.77	26.66	33.95
Fig. 7(h)	SSIM	0.8119	0.8511	0.8543	0.8584	0.8587	0.8558	0.9735

$f (x)$	original scene
${\hat{f}}_{W} (x)$	super-resolved version of $g_{W}^{I} (x)$
${\hat{f}}_{W}^{C} (x)$	super-resolved central region of $g_{W}^{I} (x)$
${\hat{f}}_{W}^{B} (x)$	super-resolved boundary region of $g_{W}^{I} (x)$
$G$	geometric transformation from $g_{W}^{I} (x)$ to $g_{T} (x)$ , such as $g_{T} (x) = G g_{W}^{I} (x)$
$g_{W} (x)$	input wide-view image
$g_{T} (x)$	input tele-view image
$g_{W}^{I} (x)$	interpolated and then cropped wide-view image
$g_{T}^{R} (x)$	registered tele-view image
$g_{W}^{D} (x)$	deconvoled result of $g_{W}^{I} (x)$
$g_{T}^{H} (x)$	high-frequency component of the registered tele-view image
$g_{T}^{L} (x)$	low-frequency component of the registered tele-view image
$g_{W}^{H} (x)$	high-frequency component of the input wide-view image
$g_{W}^{L} (x)$	low-frequency component of the input wide-view image
$H_{L} (\cdot)$	geometric degradation operator by the wide-angle lens
$H_{S} (\cdot)$	subsampling operator
$H_{C} (\cdot)$	cropping operator to model the optical zooming
$H_{I} (\cdot)$	intensity transformation operator representing the relationship between the wide-view and tele-view images
$l_{W}$	focal length of the wide-view camera
$l_{T}$	focal length of the tele-view camera
$η_{W} (x)$	additive noise in the wide-view image
$η_{T} (x)$	additive noise in the tele-view image
$p_{C} (f, x_{i})$	patch in the center region of an image $f$ centered at $x_{i}$
$p_{B} (f, x_{i})$	patch in the boundary region of an image $f$ centered at $x_{i}$
$x$	two-dimensional pixel coordinate vector, $x = {[x y]}^{T}$
$Ω_{i}$	patch region centered at $x_{i}$
$σ_{B}$	standard deviation of the Gaussian low-pass filter

Image	Method	[21]	[26]	[9]	[11]	[27]	[28]	Proposed
Fig. 7(a)	PSNR	26.22	27.72	28.25	29.06	28.65	29.15	30.13
Fig. 7(a)	SSIM	0.9603	0.9736	0.9753	0.9799	0.9780	0.9803	0.9834
Fig. 7(b)	PSNR	33.37	34.13	34.50	34.80	35.02	34.68	36.52
Fig. 7(b)	SSIM	0.9501	0.9596	0.9611	0.9645	0.9644	0.9640	0.9757
Fig. 7(c)	PSNR	27.80	29.06	29.04	29.47	29.52	29.79	30.02
Fig. 7(c)	SSIM	0.9531	0.9623	0.9615	0.9637	0.9641	0.9650	0.9789
Fig. 7(d)	PSNR	29.31	30.02	30.12	30.78	30.74	30.89	31.18
Fig. 7(d)	SSIM	0.9771	0.9817	0.9817	0.9841	0.9838	0.9847	0.9865
Fig. 7(e)	PSNR	28.05	29.72	30.06	30.80	30.71	31.25	31.45
Fig. 7(e)	SSIM	0.9634	0.9757	0.9746	0.9782	0.9776	0.9791	0.9825
Fig. 7(f)	PSNR	28.15	29.54	29.03	30.02	29.95	30.34	32.22
Fig. 7(f)	SSIM	0.9255	0.9453	0.9422	0.9515	0.9507	0.9526	0.9666
Fig. 7(g)	PSNR	31.38	33.11	32.87	33.78	33.91	33.79	33.94
Fig. 7(g)	SSIM	0.9819	0.9883	0.9867	0.9899	0.9901	0.9896	0.9900
Fig. 7(h)	PSNR	30.02	30.78	30.93	31.48	31.58	31.35	33.06
Fig. 7(h)	SSIM	0.9138	0.9326	0.9338	0.9418	0.9415	0.9394	0.9525

Image	Method	[21]	[26]	[9]	[11]	[27]	[28]	Proposed
Fig. 7(a)	PSNR	22.93	24.70	24.80	25.11	24.92	25.42	30.55
Fig. 7(a)	SSIM	0.9164	0.9463	0.9469	0.9504	0.9486	0.9537	0.9863
Fig. 7(b)	PSNR	29.12	30.03	30.20	30.30	30.32	30.34	37.88
Fig. 7(b)	SSIM	0.8775	0.9023	0.9054	0.9061	0.9051	0.9067	0.9874
Fig. 7(c)	PSNR	24.50	26.17	25.95	26.18	26.00	26.55	29.52
Fig. 7(c)	SSIM	0.8725	0.8950	0.8941	0.8956	0.8941	0.8993	0.9443
Fig. 7(d)	PSNR	25.48	26.29	26.44	26.57	26.54	26.70	32.87
Fig. 7(d)	SSIM	0.9405	0.9537	0.9545	0.9555	0.9550	0.9571	0.9908
Fig. 7(e)	PSNR	24.61	26.64	26.40	26.70	26.48	27.16	33.14
Fig. 7(e)	SSIM	0.9209	0.9489	0.9455	0.9496	0.9469	0.9519	0.9881
Fig. 7(f)	PSNR	23.67	25.38	24.97	25.24	25.17	25.66	31.87
Fig. 7(f)	SSIM	0.8062	0.8640	0.8587	0.8654	0.8637	0.8708	0.9738
Fig. 7(g)	PSNR	28.52	30.35	29.93	30.28	30.38	30.38	35.09
Fig. 7(g)	SSIM	0.9650	0.9772	0.9746	0.9768	0.9776	0.9766	0.9920
Fig. 7(h)	PSNR	25.70	26.41	26.56	26.72	26.77	26.66	33.95
Fig. 7(h)	SSIM	0.8119	0.8511	0.8543	0.8584	0.8587	0.8558	0.9735

Abstract

Cited By

Figures (17)

Tables (3)

Equations (9)

Journal of the Optical Society of America A