(JBE Vol. 19, No. 5, September 2014) (Regular Paper) 19 5, 2014 9 (JBE Vol. 19, No. 5, September 2014) http://dx.doi.org/10.5909/jbe.2014.19.5.656 ISSN 2287-9137 (Online) ISSN 1226-7953 (Print) a) Reduction of Block Artifacts in Haze Image and Evaluation using Disparity Map Oh-Seol Kwon a).... (HRMF) (EM).. Abstract In the case of a haze image, transferring the information of the original image is difficult as the contrast leans toward bright regions. Thus, dehazing algorithms have become an important area of study. Normally, since it is hard to obtain a haze-free image, the output image is qualitatively analyzed to test the performance of an algorithm. However, this paper proposes a quantitative error comparison based on reproducing the haze image using a disparity map. In addition, a Hidden Random Markov Model and EM algorithm are used to remove any block artifacts. The performance of the proposed algorithm is confirmed using a variety of synthetic and natural images. Keyword : Haze removal, image enhancement, disparity map. a) (Department of Control & Instrumentation Engineering, Changwon National University) Corresponding Author : (Oh-Seol Kwon) E-mail: osk1@changwon.ac.kr Tel: +82-55-213-3669 2012 () (No.2012R1A1A1019461) 2014 2015. Manuscript received June, 23 2014 Revised August 11, 2014 Accepted August 26, 2014....
: (Oh-Seol Kwon : Reduction of Block Artifacts of Haze Image and Evaluation using Disparity Map) [1-2].. Tan [3] (contrast)., (halo effect). Fattal [4].. ICA(Independent Component Analysis).. He Dark Channel Prior(DCP) [5]. 0.,. He. He Guided Filtering [6]....... DCP.,., (transmission). (1) R,G,B. Dark Channel Prior(DCP). DCP, 0. min min (2) x. 0.,. A. (1) min
(JBE Vol. 19, No. 5, September 2014) min min A. 1 DCP. min min, rgb min. min min min min (2) 0. min min. (a) 1. ; (a), (b) DCP Fig. 1. Example of block artfiact; (a) haze image and (b) DCP method (b) III. HMRF-EM HMRF-EM. DCP..,. HMRF [7]. arg, L. Gibbs,. EM., HMRF-EM 2.,.
: (Oh-Seol Kwon : Reduction of Block Artifacts of Haze Image and Evaluation using Disparity Map).. (a) (b) HMRF-EM...., 0. (c) 2. ; (a), (b), (c) ( ), (c) ( ) Fig. 2. Creation of haze image using depth map; (a) haze-free image, (b) disparity map, (c) haze image( ), and (d) haze image ( ) (d) arg... 2.. Hirschmller et al. [8].. 2(c) 2(d).. 3 DCP (a) (c) 3. ; (a) DCP, (b), (c) (a), (d) (b) Fig. 3. Comparison of block artfiact reduction; (a) DCP method, (b) proposed method, (c) zooming of (a), and (d) zooming of (b). (b) (d)
(JBE Vol. 19, No. 5, September 2014). 4. 5. 5. DCP, Guided. DCP Guided. RGB 1. (a) (b) (c) (d) 4. (a) (b) DCP (c) Guided (d) Fig. 4. Comparison of result images; (a) input image (b) DCP method (c) Guided filtering, and (d) proposed method.
: (Oh-Seol Kwon : Reduction of Block Artifacts of Haze Image and Evaluation using Disparity Map) (a) (b) (c) 5. ; (a) DCP (c) Guided (d) Fig. 5. Comparison of inverted error images; (a) DCP method (c) Guided filtering, and (d) proposed method 1. Table 1. Quantitative comparison of three algorithms DCP method Guided filtering Proposed Image 1 5.2959 5.2960 5.2836 Image 2 4.7651 4.7434 4.7044 Image 3 4.4898 4.3766 4.3384 Image 4 7.9433 7.9925 7.9968 4. 6 2.,
(JBE Vol. 19, No. 5, September 2014) (a) (b) (c) 6. (a) DCP (b) Guided (c) Fig. 6. Comparison of transmission maps; (a) DCP method (b) Guided filtering, and (c) proposed method...,,. 7. Fattal,
: (Oh-Seol Kwon : Reduction of Block Artifacts of Haze Image and Evaluation using Disparity Map) 7. (a), (b) Fattal (c) Guided (d) Fig. 7. Dehazing results for natural images (a) input image, (b) Fattal s method, (c) Guided filtering, and (d) proposed method 2. Table 2. Quantitative comparison of transmission errors Transmission error (J) DCP method Guided filtering Proposed 35.85 10-3 35.80 10-3 35.87 10-3.. DCP HMRF-EM...
(JBE Vol. 19, No. 5, September 2014),. (References) [1] S. G. Narasimhan and S. K. Nayar, Contrast restoration of weather degraded image, IEEE trans. Pattern Anal. Mach. Intell., vol. 25, no. 6, pp. 713-724, June 2003 [2] S. K. Nayar and S. G. Narasimhan, Vision in bad weather, Proc. ICCV, pp. 820-827, Kerkyra, Greece, Sep. 1999 [3] R. Tan, Visibility in bad weather from a single image, Proc. CVPR, pp. 1-8, Anchorage, USA, June 2008 [4] R. Fattal, Single image dehazing, ACM Trans. Graphics, vol. 27, no. 3, pp. 1-9, Aug. 2008 [5] 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. 23412353, Oct. 2011 [6] K. He, J. sun, and X. Tang, Guided imaging filtering," IEEE trans. Pattern Anal. Mach. Intell., vol. 35, no. 6, pp. 1397-1409, June 2013 [7] Q. Wang, "HMRF-EM-image: Implementation of the hidden markov random field and its expectation-maximization algorithm," arxiv: 1207. 3510 [cs.cv], 2012 [8] H. Hirschm and D. Scharstein, "Evaluation of cost functions for stereo matching," Proc. CVPR, pp. 1-8, Minneapolis, USA, 2007-2008 8 : - 2008 9 ~ 2010 8 : New York University - 2010 9 ~ 2011 8 : - 2011 9 ~ : - :,,