GA-Neural Approach for Latent Finger Print Matching
GA-Neural Approach for Latent Finger Print Matching
复制标题
DOI：
10.1109/isms.2011.19
复制
发表时间：
2011-01
期刊：
影响因子：
--
通讯作者：
Shahrzad Shapoori;N. Allinson
中科院分区：
文献类型：
--
作者：
Shahrzad Shapoori;N. Allinson
潜在指纹匹配是科学中最新鲜的领域之一。目前的潜在指纹匹配方法是人工的，依赖于人的经验。遗憾的是，目前还没有一种能够自动进行潜在指纹匹配的系统。眼动追踪技术能够记录用户的眼球运动，为用户的搜索策略提供有用的信息。本文对眼动仪采集的实验数据进行聚类分析，并设计了一个基于神经网络的专家搜索策略学习系统。结果表明，该系统能够根据专家经验预测最优搜索策略。
Latent finger print matching is one of the freshest areas in science. The current methods of latent finger print matching are manual and reliable on human experience. Unfortunately, a system, which can perform the latent fingerprint matching automatically, does not exist. The eye tracking technology is able to record the eye movement and could provide useful information about the user search strategy. In this paper, the experimental data obtained from an eye tracker is analyzed by clustering analysis and a neural network based system is designed to learn the search strategy of the experts. The results show that the system is able to predict the optimum search strategy based on expert’s experiences.