Adaptive Weighted Generalized Total Variation Image Deblurring Based on Primal-Dual algorithm

Aiping Yang; Yue Zhang; Jinbin Wang; Yuqing He

doi:10.3788/LOP55.041003

Journals >Laser & Optoelectronics Progress >Volume 55 >Issue 4 >Page 041003 > Article

Laser & Optoelectronics Progress
Vol. 55, Issue 4, 041003 (2018)

Adaptive Weighted Generalized Total Variation Image Deblurring Based on Primal-Dual algorithm

Aiping Yang, Yue Zhang, Jinbin Wang^*, and Yuqing He

Author Affiliations

School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China

show less

DOI: 10.3788/LOP55.041003 Cite this Article Set citation alerts

Aiping Yang, Yue Zhang, Jinbin Wang, Yuqing He. Adaptive Weighted Generalized Total Variation Image Deblurring Based on Primal-Dual algorithm[J]. Laser & Optoelectronics Progress, 2018, 55(4): 041003 Copy Citation Text

show less

Comparison of deblurred images. (a) Blurred image; (b) result of FTVD algorithm; (c) result of TV-NLR algorithm; (d) result of JSM algorithm; (e) result of the proposed algorithm; (f)-(i) detail enlarged drawing of Fig. (b)-(e)

Fig. 1. Comparison of deblurred images. (a) Blurred image; (b) result of FTVD algorithm; (c) result of TV-NLR algorithm; (d) result of JSM algorithm; (e) result of the proposed algorithm; (f)-(i) detail enlarged drawing of Fig. (b)-(e)

Download full size

Comparison of deblurred images. (a) Blurred image; (b) result of FTVD algorithm; (c) result of TV-NLR algorithm; (d) result of JSM algorithm; (e) result of he proposed algorithm; (f)-(i)detail enlarged drawing of Fig. (b)-(e)

Fig. 2. Comparison of deblurred images. (a) Blurred image; (b) result of FTVD algorithm; (c) result of TV-NLR algorithm; (d) result of JSM algorithm; (e) result of he proposed algorithm; (f)-(i)detail enlarged drawing of Fig. (b)-(e)

Download full size

Fig. 3. Comparison of deblurred images. (a) Blurred image; (b) result of FTVD algorithm; (c) result of TV-NLR algorithm; (d) result of JSM algorithm; (e) result of the proposed algorithm; (f)-(i) detail enlarged drawing of Fig. (b)-(e)

Download full size

Algorithm	House	Lena	Barbara	Model	Yacht
Union blur kernel, 9×9 σ=0.005
FTVD	36.06	34.57	26.67	39.62	36.04
TV-NLR	36.09	35.53	24.35	40.76	36.17
JSM	36.54	34.25	27.01	39.88	36.54
Proposed	47.87	44.74	39.55	51.05	49.33
Gaussian blur kernel: (25, 1.6), σ=0.005
FTVD	34.70	34.96	25.82	40.07	34.90
TV-NLR	34.82	36.11	23.07	40.67	35.32
JSM	35.38	35.57	25.94	41.75	36.95
Proposed	38.71	39.01	29.70	45.81	41.07
Motion blur kernel: (20, 45), σ=0.005
FTVD	36.37	36.20	29.82	40.39	36.12
TV-NLR	36.15	36.24	32.66	40.53	36.18
JSM	36.39	36.07	31.34	41.00	36.87
Proposed	50.89	49.81	47.60	54.52	52.96

Table 1. PSNR results comparison of four deblurred imagesdB

Algorithm	House	Lena	Barbara	Model	Yacht
Union blur kernel, 9×9 σ=0.005
FTVD	3.32	12.34	30.78	7.35	14.28
TV-NLR	5.22	35.53	45.23	42.11	42.07
JSM	160.77	653.98	688.07	667.13	678.09
Proposed	1.65	7.18	7.28	7.09	7.39
Gaussian blur kernel: (25, 1.6), σ=0.005
FTVD	1.78	7.78	25.29	4.81	10.92
TV-NLR	14.85	62.42	64.11	59.17	58.76
JSM	177.57	689.1	715.12	692.66	702.79
Proposed	1.65	7.17	7.24	7.30	7.29
Motion blur kernel: (20, 45), σ=0.005
FTVD	3.60	13.17	39.32	7.56	15.57
TV-NLR	2.37	23.73	20.60	20.87	37.46
JSM	176.68	722.444	760.95	703.84	725.00
Proposed	1.65	7.18	7.34	7.34	7.30

Table 2. Time complexity comparison of four algorithmss

Aiping Yang, Yue Zhang, Jinbin Wang, Yuqing He. Adaptive Weighted Generalized Total Variation Image Deblurring Based on Primal-Dual algorithm[J]. Laser & Optoelectronics Progress, 2018, 55(4): 041003

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information