ABC-SVM模型在手指静脉图像质量评估中的应用研究  

The Application Research of the ABC-SVM Model in the Finger Vein Image Quality Assessment

作　　者：胡晶晶 秦华锋[2] 罗钦 周树林 曾建梅 李翠锦 HU Jing jing QIN Hua-feng LUO Qin ZHOU Shu-lin ZENG Jian-mei LI Cui-jin(Chongqing Institute of Engineering, Chongqing 400056, China Chongqing Technology and Business University, Chongqing 400056, Chin)

机构地区：[1]重庆工程学院,重庆400056 [2]重庆工商大学,重庆400060

出　　处：《激光杂志》2017年第4期169-172,共4页Laser Journal

基　　金：国家自然科学基金资助项目(61402063);重庆市教委科学技术研究项目资助(KJ1601704);重庆工程学院校内基金资助(2015xzky06)

摘　　要：为了避免传统优化方法容易陷入局部最优解的情况,本文采用人工蜂群(Artificial Bee Colony,ABC)方法对支持向量机(Support Vector Machine,SVM)模型的惩罚因子C和宽度参数σ进行参数优化,兼顾了局部最优解和全局最优解,建立了ABC-SVM模型,并将此模型应用在手指静脉图像质量评估中。通过与未经过参数优化的SVM模型对比,同时也与蚁群算法(ACO)、遗传算法(GA)、粒子群算法(PSO)三种优化方法进行对比实验,实验结果表明,ABC-SVM模型无论在分类准确率方面,还是运行时间方面,都是可行的,证明其具有良好的应用价值。In order to avoid falling into local optimal solution of the traditional optimization methods, the penalty factor C and width parameter σ of the support vector machine model are optimized by the artificial bee colony method, with which the global optimal solutions and local optimal solutions are combined, and the ABC-SVM model is estab- lished, then the model is applied in the finger vein image quality assessment. Compared with the SVM model, which is not optimized, also with the three optimization methods of ant colony algorithm, genetic algorithm and particle swarm optimization, the experimental results prove that the ABC-SVM model is feasible in terms of the classification accuracy and the running time, It has good application value.

关 键 词：人工蜂群方法 支持向量机 参数优化 质量评估 

分 类 号：TN911[电子电信—通信与信息系统]