基于视觉感光层功能的菌落图像多强度边缘检测研究

doi:10.3969/j.issn.0258-8021. 2014. 06.07

摘要
图/表
参考文献
相关文章 (2)

全文: PDF (2179 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要借鉴视觉神经系统在轮廓感知中的独特优势，提出一种基于视觉感光层功能的图像边缘检测新方法。构建以带漏感的积累发放（LIF）神经元电生理模型为基本单元的神经元网络；根据特定时间窗口内各个神经元的脉冲发放情况，对神经元的增强(ON)或抑制(OFF)类别进行判断；通过拮抗式感受野特性以及神经元激励的反馈增强模式，实现弱边缘的凸显；为克服视觉感光层所具有的适应性并凸显弱细节的对比度，对图像进行多方向、多距离尺度的移动，并融合感光层神经元网络脉冲发放率的差异信息，最后实现图像边缘的有效检测。以具有丰富边缘特性的20幅菌落图像为样本，以边缘置信度和重构相似度作为评价指标，对多强度边缘进行检测。结果表明，所提出方法可以有效完整地检测出图像多强度边缘，且其对弱边缘检测的重构相似度均值高于08，检测准确性有显著的提高(P<005)。所提出的利用生理视觉系统特性进行边缘检测，为包含多强度边缘信息的图像处理提供崭新的思路。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	罗佳骏武薇^*范影乐高云园

关键词 ：边缘检测, 感光层, 感受野, LIF神经元模型

Abstract：Considering the superiority of visual nervous system in contour perception, a new method of image edge detection based on the function of photoreceptor in visual system was proposed in this paper. Firstly, the neural network was constructed with leaky integrateandfire (LIF) neuronal electrophysiological model. Secondly, each neuron would be classified as excitation (ON) type or inhibition (OFF) type according to the neural firing pattern. And then the weak edges were highlighted by using centersurround antagonistic receptive field feature and feedback enhancing mode of neuronal excitation. Meanwhile the image movement in multidirection and multiscale was applied to overcome the adaptability of photoreceptor and highlight the contrast of weak details. Finally the edge image was acquired by fusing the variance information, such as photosensitive neural network’s firing rates. In this paper, twenty colony images having rich details in the edge were selected as experimental samples. The results of multiintensity edge detection were assessed by the confidence of edge and reconstruction similarity. It was proved that the new method can effectively detect the intact multiintensity edge image, especially the means of reconstruction similarity in the weak edge detection is improved significantly, which is higher than 08 (P<005). The method of edge detection discussed in this paper provides a brandnew idea for multiintensity edge details image processing based on the feature of physiological visual.

Key words： edge detection photoreceptor receptive field leaky integrate\ and\ fire (LIF) neuronal model

基金资助:国家自然科学基金(60872090, 61201300)；浙江省教育厅科研项目(Y201328513)

引用本文:

罗佳骏武薇^*范影乐高云园. 基于视觉感光层功能的菌落图像多强度边缘检测研究[J]. 中国生物医学工程学报, 2014, 33(6): 677-686.
LUO Jia Jun WU Wei^* FAN Ying Le GAO Yun Yuan. Multiintensity Edge Detection for Colony Images Based on the Function of Photoreceptor in Visual System #br#. journal1, 2014, 33(6): 677-686.

链接本文:

http://cjbme.csbme.org/CN/10.3969/j.issn.0258-8021. 2014. 06.07 或 http://cjbme.csbme.org/CN/Y2014/V33/I6/677

［1］Wenshuo G, Xiaoguang Z, Lei Y, et al. An improved Sobel edge detection ［C］// 3rd IEEE International Conference on Computer Science and Information Technology (ICCSIT). Chengdu: IEEE, 2010(5): 67-71.
［2］Bing W, Shaosheng F. An improved CANNY edge detection algorithm ［C］ // Qingling L, Fei Y, Yun L， et al. eds. Second IEEE International Workshop on Computer Science and Engineering. Qingdao: IEEE, 2009: 497-500.
［3］Zhan S, Jinglu H. Image edge detection method based on a simplified PCNN model with anisotropic linking mechanism ［C］ // Aboul EH, Ajith A, Francesco M, et al. eds. 10th IEEE International Conference on Intelligent Systems Design and Applications (ISDA). Cairo: IEEE, 2010:330-335.
［4］Yiming J, Chang HK, ChiCheng H. Efficient edge detection and object segmentation using Gabor filters ［C］//Proceedings of the 42nd Annual Southeast Regional Conference. Huntsville: ACM, 2004: 454-459.
［5］Lee BB, Martin PR, Grünert U. Retinal connectivity and primate vision ［J］. Prog Retin Eye Res, 2010, 29(6): 622-639.
［6］Mohemmed A, Schliebs S, Matsuda S, et al. Training spiking neural networks to associate spatiotemporal input-output spike patterns ［J］. Neurocomputing, 2013, 107: 3-10.
［7］Chouhan AS. An analytical study of leaky integrate-andfire neuron model using MATLAB simulation ［J］. Int J Eng Res Technol, 2013, 2(4): 2242-2245.
［8］Pillow JW, Shlens J, Paninski L, et al. Spatiotemporal correlations and visual signaling in a complete neuronal population ［J］. Nature, 2008, 454(7207): 995-999.
［9］Tanabe S. Population codes in the visual cortex ［J］. Neurosci Res, 2013, 76(3): 101-105.
［10］MartinezConde S, Macknik SL, Troncoso XG, et al. Microsaccades counteract visual fading during fixation ［J］. Neuron, 2006, 49(2): 297-305.
［11］Kuang X, Poletti M, Victor JD, et al. Temporal encoding of spatial information during active visual fixation ［J］. Curr Biol, 2012, 22(6): 510-514.
［12］Meer P, Georgescu B. Edge detection with embedded confidence ［J］. IEEE Trans Pattern Anal Mach Intell, 2001, 23(12): 1351-1365.
［13］Wang Z, Bovik AC, Sheikh HR, et al. Image quality assessment: from error visibility to structural similarity ［J］. IEEE Trans Image Process, 2004, 13(4): 600-612.
［14］Govindarajani B, Panetta KA, Agaian S. Image reconstruction for quality assessment of edge detectors ［C］// IEEE International Conference on Systems, Man and Cybernetics. Singapore: IEEE, 2008: 691-696.