Single Image Haze Removal Using Iterative Ambient Light Estimation with Region Segmentation

Fig. 22

Table 15

Relationship between tsu m and PSNR.

Behavior of the remained haze tsu m (Actual haze image).

Image

Tienanmen (Proposed 1)

Tienanmen (Proposed 2)

Snow (Proposed 1)

Snow (Proposed 2)

5.3

0.84

0.94

0.71

0.86

number of iteration

0.88

0.96

0.83

0.95

0.91

0.97

0.91

0.97

0.91

0.97

0.96

0.99

Case of Images with Actual Haze

Table 15 shows the behavior of the remaining haze t sum

in each iteration for images with actual haze (i.e. images

without a ground truth). Again, the numbers in bold in

Table 15 indicate when the iteration terminated. Similarly to

Tables 13 and 14, we see from Table 15 that the remaining

haze t sum increases as the number of iterations increases.

According to the difference of t sum for each iteration, it

decreases as the number of iterations increases. This can

be considered as convergence, also confirmed in Tables 13

and 14. Again, the haze is well removed after the estimated

number of iterations because the remaining haze t sum when

the iteration is terminated again exceeds 0.7 and approaches

one.

[1] G.A. Woodell, D.J. Jobson, Z.-U. Rahman, and G. Hines, “Advanced

image processing of aerial imagery,” Proc. SPIE, vol.6246, p.62460E,

May 2006.

[2] Y. Luo, T. Liu, D. Tao, and C. Xu, “Decomposition-based transfer

distance metric learning for image classification” IEEE Trans. Image

Process., vol.23, no.9, pp.3789–3801, Sept. 2014.

[3] J. Han, D. Zhang, G. Cheng, L. Guo, and J. Ren, “Object detection in

optical remote sensing images based on weakly supervised learning

and high-level feature learning,” IEEE Trans. Geosci. Remote Sens.,

vol.53, no.6, pp.3325–3337, June 2015.

[4] L. Liu and L. Shao, “Learning discriminative representations from

RGB-D video data,” Proc. Int. Joint Conf. Artif. Intell., pp.1493–

1500, Aug. 2013.

[5] R.T. Tan, “Visibility in bad weather from a single image,” Proc. IEEE

Conf. Comput. Vis. Pattern Recognit. (CVPR), no.1063-6919, June

2008.

[6] K. Tang, J. Yang, and J. Wang, “Investigating haze-relevant features

in a learning framework for image dehazing,” Proc. IEEE Conf.

Comput. Vis. Pattern Recognit. (CVPR), pp.2995–3002, June 2014.

[7] R. Fattal, “Single image dehazing,” ACM Trans. Graph., vol.27, no.3,

p.72, Aug. 2008.

[8] R. Fattal, “Dehazing using color-lines,” ACM Trans. Graphics,

vol.34, no.1, article no.13, Dec. 2014.

[9] M. Sulami, I. Geltzer, R. Fattal, and M. Werman, “Automatic recovery of the atmospheric light in hazy images,” Proc. Int. Conf.

Computational Photography (ICCP), pp.1–11, May 2014.

[10] K. He, J. Sun, and X. Tang, “Single image haze removal using

dark channel prior,” IEEE Trans. Pattern Anal. Mach. Intell., vol.33,

no.12, pp.2341–2353, Dec. 2011.

[11] Y. Cheng, B. Chen, S. Huang, S. Kuo, A. Kopylov, O. Seredint, L.

Mestetskiy, B. Vishnyakov, Y. Viziter, O. Vygolov, C. Lian, and C.

Wu, “Visibility enhancement of single hazy images using hybrid dark

channel prior,” IEEE International Conference on Systems, Man, and

Cybernetics, pp.3627–3632, Oct. 2013.

[12] B. Cai, X. Xu, K, Jia, C. Qing, and D. Tao, “DehazeNet: An endto-end system for single image haze removal,” IEEE Trans. Image

Process., vol.25, no.11, pp.5187–5198, Nov. 2016.

[13] H. Zhang and V.M. Patel, “Densely connected pyramid dehazing

network,” Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR),

June 2018.

[14] Q. Zhu, J. Mai, and L, Shao, “A fast single image haze removal

IEICE TRANS. FUNDAMENTALS, VOL.E104–A, NO.2 FEBRUARY 2021

562

[15]

[16]

[17]

[18]

[19]

[20]

[21]

algorithm using color attenuation prior,” IEEE Trans. Image Process.,

vol.24, no.11, pp.3522–3533, Nov. 2015.

K. He, J. Sun, and X. Tang, “Guided image filtering “, IEEE Trans.

Pattern Anal. Mach. Intell., vol.35, no.6, pp.1397–1409, June 2013.

H. Yuan, C. Liu, Z, Guo, and Z. Sun, “A region-wised medium

transmission based image dehazing method,” IEEE Access, vol.5,

pp.1735–1742, Jan. 2017.

D. Comaniciu and P. Meer, “Mean shift: A robust approach toward

feature space analysis,” IEEE Trans. Pattern Anal. Mach. Intell.,

vol.24, no.5, pp.603–619, May 2002.

G. Strang, Introduction to Linear Algebra, 4th ed., WellesleyCambridge Press, USA, 2009.

Y. Araki and K. Ichige, “Single image haze removal using ambient

light estimation and region segmentation,” Proc. Int. Sympo. Intelligent Signal Processing and Communication Systems, pp.70–75, Nov.

2018.

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., vol.28, no.1, pp.492–505, Jan. 2019.

C.O. Ancuti, C. Ancuti, R. Timofte, and C. De Vleeschouwer, “OHAZE: A dehazing benchmark with real hazy and haze-free outdoor images,” Proc. IEEE/CVF Conf. Computer Vision and Pattern

Recognition Workshops (CVPRW), June 2018.

Yuji Araki

received B.E. and M.E. degrees in Electrical and Computer Engineering

from Yokohama National University in 2017 and

2019, respectively. Currently he is with NEC

Corporation. During his masters’ study, he engaged with the research of digital image processing techniques like restoration or denoising. He

received Best Paper Award in Int. Sympo. Intelligent Signal Processing and Communication

Systems (ISPACS) in 2018.

Kentaro Mita

received B.E. and M.E. degrees in Electrical and Computer Engineering

from Yokohama National University in 2018 and

2020, respectively. Currently he is with Yokogawa Electric Corporation. During his masters’

study, he engaged with the research of machine

learning approach to digital image processing

techniques like restoration or separation.

Koichi Ichige

received B.E., M.E. and

Dr. Eng. degrees in Electronics and Computer

Engineering from the University of Tsukuba in

1994, 1996 and 1999, respectively. He joined

the Department of Electrical and Computer Engineering, Yokohama National University as a

research associate in 1999, where he is currently

a professor. He has been on leave to Swiss Federal Institute of Technology Lausanne (EPFL),

Switzerland as a visiting researcher in 2001–

2002. His research interests include digital signal processing, approximation theory and their applications to image processing and mobile communication. He served as an associate editor of

IEEE Transactions on Industrial Electronics in 2004–2008, Journal of Circuits, Systems and Computers (JCSC) in 2012–2014, and IEICE Transactions on Fundamentals of Electronics, Communications and Computer

Sciences (IEICE-EA) in 2015–2018. Currently he serves as an editor of

IEICE-EA. He received “Meritorious Award on Radio” from the Association of Radio Industries and Businesses (ARIB) in 2006, Best Letter Award

from IEICE Communication Society in 2007, and Best Paper Award in

Int. Sympo. Intelligent Signal Processing and Communication Systems

(ISPACS) in 2018. He is a member of IEEE.

...