Saliency Motivated Pulse Coupled Neural Network for Underwater Laser Image Segmentation

doi:10.1007/s12204-016-1724-1

摘要/Abstract

摘要： The detection range of underwater laser imaging technology achieves 4—6 times of detection range of conventional camera in intervening water medium, which makes it very promising in oceanic research, deep sea exploration and robotic works. However, the special features in underwater laser images, such as speckle noise and non-uniform illumination, bring great difficulty for image segmentation. In this paper, a novel saliency motivated pulse coupled neural network (SM-PCNN) is proposed for underwater laser image segmentation. The pixel saliency is used as external stimulus of neurons. For improvement of convergence speed to optimal segmentation, a gradient descent method based on maximum two-dimensional Renyi entropy criterion is utilized to determine the dynamic threshold. On the basis of region contrast in each iteration step, the real object regions are effectively distinguished, and the robustness against speckle noise and non-uniform illumination is improved by region selection. The proposed method is compared with four other state-of-the-art methods which are watershed, fuzzy C-means, meanshift and normalized cut methods. Experimental results demonstrate the superiority of our proposed method to allow more accurate segmentation and higher robustness.

关键词: underwater laser image, pulse coupled neural network, pixel saliency, region contrast

Abstract: The detection range of underwater laser imaging technology achieves 4—6 times of detection range of conventional camera in intervening water medium, which makes it very promising in oceanic research, deep sea exploration and robotic works. However, the special features in underwater laser images, such as speckle noise and non-uniform illumination, bring great difficulty for image segmentation. In this paper, a novel saliency motivated pulse coupled neural network (SM-PCNN) is proposed for underwater laser image segmentation. The pixel saliency is used as external stimulus of neurons. For improvement of convergence speed to optimal segmentation, a gradient descent method based on maximum two-dimensional Renyi entropy criterion is utilized to determine the dynamic threshold. On the basis of region contrast in each iteration step, the real object regions are effectively distinguished, and the robustness against speckle noise and non-uniform illumination is improved by region selection. The proposed method is compared with four other state-of-the-art methods which are watershed, fuzzy C-means, meanshift and normalized cut methods. Experimental results demonstrate the superiority of our proposed method to allow more accurate segmentation and higher robustness.

Key words: underwater laser image, pulse coupled neural network, pixel saliency, region contrast

中图分类号:

TP 183

WANG Bo* (王博), WAN Lei (万磊), LI Ye (李晔). Saliency Motivated Pulse Coupled Neural Network for Underwater Laser Image Segmentation[J]. 上海交通大学学报（英文版）, 2016, 21(3): 289-296.

WANG Bo* (王博), WAN Lei (万磊), LI Ye (李晔). Saliency Motivated Pulse Coupled Neural Network for Underwater Laser Image Segmentation[J]. Journal of shanghai Jiaotong University (Science), 2016, 21(3): 289-296.

参考文献 15

[1]	TULLDAHL H M, ANDERSON P, OLSSON A, etal. Experimental evaluation of underwater range-gatedviewing in natural waters [J]. Proceedings of SPIE,2006, 6395: 639506.
[2]	GE W L, ZHANG X H. Design and implementationof range-gated underwater laser imaging system [J].Proceedings of SPIE, 2014, 9142: 914216.
[3]	OUYANG B, DALGLEISH F R, CAIMI F M, et al.Image enhancement for underwater pulsed laser linescan imaging system [J]. Proceedings of SPIE, 2012,8372: 83720R-1.
[4]	HUANG Y W, CAO F M, JIN W Q, et al. Underwaterpulsed laser range-gated imaging model and its effecton image degradation and restoration [J]. Optical Engineering,2014, 53(6): 061608.
[5]	JOHNSON J L, PADGETT M L. PCNN models andapplications [J]. IEEE Transactions on Neural Networks,1999, 10(3): 480-498.
[6]	RANGANATH H S, KUNTIMAD G. Object detectionusing pulse coupled neural networks [J]. IEEE Transactionson Neural Networks, 1999, 10(3): 615-620.
[7]	BERG H, OLSSON R, LINDBLAD T, et al. Automaticdesign of pulse coupled neurons for image segmentation[J]. Neurocomputing, 2008, 71(10): 1980-1993.
[8]	OMIDVAR O, DAYHOFF J. Neural networks and patternrecognition [M]. New York: Academic Press, 1998:1-56.
[9]	ITTI L, KOCH C, NIEBUR E. A model of saliencybasedvisual attention for rapid scene analysis [J].IEEE Transactions on Pattern Analysis and MachineIntelligence, 1998, 20(11): 1254-1259.
[10]	ACHANTA R, HEMAMI S, ESTRADA F, et al.Frequency-tuned salient region detection [C]//IEEEConference on Computer Vision and Pattern Recognition.Miami, USA: IEEE, 2009: 1597-1604.
[11]	DENG X Y, MA Y D. PCNN model analysis and itsautomatic parameters determination in image segmentationand edge detection [J]. Chinese Journal of Electronics,2014, 23(1): 97-103.
[12]	CHENG M M, ZHANG G X, MITRA N J, et al. Globalcontrast based salient region detection [C]//IEEEConference on Computer Vision and Pattern Recognition.Colorado, USA: IEEE, 2011: 409-416.
[13]	ZHANG Y D, WU L N. Image segmentation based on2D Tsallis entropy with improved pulse coupled neuralnetworks [J]. Journal of Southeast University (NaturalScience Edition), 2008, 38(4): 579-584 (in Chinese).
[14]	MA Y D, DAI R L, LI L. Automated image segmentationusing pulse coupled neural networks and image’sentropy [J]. Journal of China Institute of Communications,2002, 23(1): 46-51 (in Chinese).
[15]	YAN Q, XU L, SHI J P, et al. Hierarchical saliencydetection [C]//IEEE Conference on Computer Visionand Pattern Recognition. Portland, USA: IEEE, 2013:1155-1162.