基于免疫克隆与核匹配追踪的快速图像目标识别

缑水平; 焦李成; 张向荣; 李阳阳

doi:10.3724/SP.J.1146.2007.01491

基于免疫克隆与核匹配追踪的快速图像目标识别

doi: 10.3724/SP.J.1146.2007.01491

基金项目:

国家十一五预研项目 (51307040103)和国家863计划项目(20060101Z1119)资助课题

计量
- 文章访问数: 3441
- HTML全文浏览量: 60
- PDF下载量: 992
- 被引次数: 0
出版历程
- 收稿日期: 2007-09-18
- 修回日期: 2007-12-24
- 刊出日期: 2008-05-19

Kernel Matching Pursuit Based on Immune Clonal Fast Algorithm for Image Object Recognition

摘要

摘要: 为了避免核匹配追踪通过贪婪算法在基函数字典中寻找一组基函数的线性组合来逼近目标函数的计算量大的缺陷,本文利用免疫克隆选择算法全局最优和局部快速收敛的特性,加快对核匹配追踪算法每次的匹配过程进行优化,提出了一种免疫克隆核匹配追踪图像目标识别算法,该算法有效降低了核匹配追踪算法的计算量,对UCI数据集和遥感图像进行的仿真实验结果表明,相比标准核匹配追踪,该算法保持相当识别率情况下可以明显缩短一次匹配追踪的时间,尤其当字典规模较大时效果更为明显;同基于遗传算法优化相比,本文方法目标识别速度快,精度高。
- 图像目标识别; 核匹配追踪; 贪婪算法; 免疫克隆选择算法; 遗传算法
Abstract: In order to avoid the default of the greedy algorithm to Approximate given function by searching a linear combination of basis functions choosing from a redundant basis function dictionary for the Kernel Matching Pursuits (KMP), we make use of the global optimal searching ability and the locally quickly searching ability of Immune Clonal Selection Algorithm (ICSA) to speed up searching basic function data in function dictionary. And a method for object recognition of Kernel matching pursuits based on immune clonal selection algorithm is presented. This method reduces greatly computer time of the KMP algorithm. The simulation result of the UCI datasets, remote images and Brodatz images show the proposed algorithm can decrease obviously training time leave the classification accuracy almost unchanged, especially for the large size datasets as compared with the standard KMP. The method has higher classification speed and more accurate recognition rate over the matching pursuits based on Genetic Algorithm (GA).

HTML全文

参考文献(1)

Mallat S and Zhang Z. Matching pursuit with time-frequencydictionaries[J].IEEE Trans. on Signal Processing.1993, 41(12):3397-3415[2]Bergeau F and Mallat S. Matching pursuit of images. InProceeding of IEEE-SP: Piladephia ed. PA, USA: IEEE Press.1994: 330-333.[3]Vincent P and Bengio Y. Kernel matching pursuit. MachineLearning, 2002, 48(1): 169-191.[4]Burges C J C. Geometry and invariance in kernel basedmethod. Advance in Kernel Method-Support Vector Learning.Cambridge, MA: MIT Press, 1999: 86-116.[5]高强, 张发启, 孙德明等. 遗传算法降低匹配追踪算法计算量的研究. 振动、测试与诊断, 2003, 23(3): 165-167.Gao Q, Zhang F Q, and Sun D M. Reduction in calculationamount of matching pursuit by gene algorithm. Journal ofVibration, Measurement Diagnosis, 2003, 23(3): 165-167.[6]范虹, 孟庆丰, 张优云. 用混合编码遗传算法实现匹配追踪算法. 西安交通大学学报, 2005, 39(3): 295-299.Fan H, Meng Q F, and Zhang Y Y. Matching pursuit viagenetic algorithm based on hybrid coding. Journal of XianJiaotong University, 2005, 39(3): 295-299.[7]李恒建, 尹忠科, 王建英. 基于量子遗传优化算法的图像稀疏分解. 西南交通大学学报, 2007, 42(1): 19-23.Li H J, Yin Z K, and Wang J Y. Image sparse decompositionbased on Quantum genetic algorithm. Journal of SouthwestJiaotong University, 2007, 42(1): 19-23.[8]Adelino R and Silva F D. Atomic decomposition withevolutionary pursuit[J].Digital signal Processing.2003, 13(2):317-337[9]焦李成, 杜海峰. 人工免疫系统进展与展望. 电子学报, 2003,31(9): 73-80.Jiao L C and Du H F. Development and prospect of artificialimmunity system. Acta Electronica Sinica, 2003, 31(9): 73-80.[10]焦李成, 杜海峰, 刘芳, 公茂果. 免疫优化计算、学习与识别.第一版, 北京：科学出版社, 2006: 92-116.Jiao L C, Du H F, and Liu F, et al.. Immunity OptimalComputer, Learning and Recognition. Edition 1, Beijing:Science Press, 2006: 92-116.[11]刘芳, 杨海潮. 参数可调的克隆多播路由算法. 软件学报,2005, 16(1): 145-150.Liu F and Yang H C. A clone based multicast algorithm withadjustable parameter. Journal of Software, 2005, 16(1):145-150.[12]李阳阳, 焦李成. 求解SAT 问题的量子免疫克隆算法. 计算机学报, 2007, 30(2): 176-183.Li Y Y and Jiao L C. Quantum-inspired immune clonalalgorithm for SAT problem. Chinese Journal of Computers,2007, 30(2): 176-183.[13]Jiao L C and Li Q. Kernel Matching Pursuit ClassifierEnsemble[J].Pattern Recognition.2006, 39(4):587-594[14]廖斌, 许刚, 王裕国. 基于非抽样小波字典的低速率视频编码.软件学报, 2004, 15(2): 221-228.Liao B, Xu G, and Wang Y G. Low bit-rate video codingbased on undecimated wavelet dictionary. Journal of Software,2004, 15(2): 221-228.[15]刘利雄, 贾云得, 廖斌等. 一种改进的最佳时频原子搜索策略.中国图像图形学报, 2004, 9(7): 873-877.Liu L X, Jia Y D, and Liao B. An improved searching schemeusing optimal time-frequency atoms. Journal of Image andGaphics, 2004, 9(7): 873-877.[16]Chang S and Carin L. Kernel matching pursuits prioritizationof wavelet coefficients for SPIHT image coding. IEEEInternational Conference on Acoustics, Speech, and SignalProcessing. Proceedings. 2004, 3(17): iii-649-652.[17]Meyer F G, and Coifman R R. Brushlets: A tool fordirectional image analysis and image compression. Appliedand Computational Harmonic Analysis, 1997, 6(4): 147-187.

施引文献

资源附件(0)

访问统计