tlyonlogo
会员中心
您好,请登录, 注册

最近浏览的商品:

0去询价篮结算
询价篮中还没有商品,赶紧选购吧!
当前位置: 首页 > 实验技术 > 离子迁移谱/质谱 > 离子迁移谱技术用于笔芯笔迹鉴定
文章分类
浏览历史

离子迁移谱技术用于笔芯笔迹鉴定


来源: Barbora ANDR´AˇSIKOV´A et al   发布日期: 2022-07-25  访问量: 34


离子迁移谱(IMS)技术在不同领域有着广泛的应用。在本文中,我们分别描述了笔制造商对笔芯类型的自动识别的概念。 ...
标签: 离子迁移谱、法医学分析、笔迹鉴定、分类算法
 

Pen Refill Recognition using Ion Mobility Spectrometry

离子迁移谱技术用于笔芯笔记鉴定

Barbora ANDR´AˇSIKOV´A, J´an BARTOˇS, Zuzana HANZL´IKOV´A, Matej
PETR´IK, Marcel VESEL´Y, Ivan VYKOPAL⋆
Slovak University of Technology in Bratislava
Faculty of Informatics and Information Technologies
Ilkovicova 2, 842 16 Bratislava, Slovakia

Abstract.

Ion Mobility Spectrometry (IMS) Technology has widespread use in different fields. In this paper, we describe our concept of automatized recognition of pen refill types, respectively, pen manufacturers.
Assuming a strict measurement procedure, we work with the hypothesis that every refill type has unique behavior measurable by an ion mobility spectrometer. As part of our research, we have used several methods for spectra classification including traditional machine learning methods such as tree classifiers, support vector classifiers, as well as deep learning
methods based on neural networks. The achieved accuracy of our models varies over 85%. The result of our work is a desktop application designed for an automatic prediction based on measured pen spectra and displaying the results of the 3 most probable classes. This concept is comprehensive enough for recognition between subjects other than pen refills if we previously use our training pipeline on new objects of interest.

离子迁移谱(IMS)技术在不同领域有着广泛的应用。在本文中,我们分别描述了笔制造商对笔芯类型的自动识别的概念。

假设有一个严格的测量程序,我们假设每种填充类型都有可通过离子迁移谱测量的独特行为。作为我们研究的一部分,我们使用了几种光谱分类方法,包括传统的机器学习方法,如树分类器、支持向量分类器以及深度学习

基于神经网络的方法。我们的模型达到的精度变化超过85%。我们的工作结果是一个桌面应用程序,用于根据测量的笔芯光谱进行自动预测,并显示3个最可能类别的结果。如果我们以前在感兴趣的新对象上使用我们的训练管道,那么这个概念足够全面,可以在除笔芯填充之外的受试者之间进行识别。


Keywords: ion mobility spectrometry · forensic analysis · pen refill · classification

关键词:离子迁移谱、法医学分析、笔芯笔迹、分类算法

...

3.1 Traditional methods
For traditional Artificial Intelligence (AI) methods, we used multiple classification methods to compare the results of each classifier to identify the best ones, which we used in creating our final application.

Data acquisition Data for our research was obtained by the Advanced Ion Mobility Spectrometer (AIMS) in collaboration with MaSaTech Our dataset contains 28 different pen refills in 168 samples. We gathered refills from 17 different manufacturers including 17 ballpoints, 10 gel-based refills and 1 oil-based refill. The original sample consists of a variable number (around 200 per sample) of spectra measured in 30-second intervals. Spectra is represented by many values of mobility (varies around 665) and the corresponding values of current.

 

 


*本文内容非商业广告,仅供专业人士参考!

相关产品
相关文章

配送方式
采购指南
付款方式
售后服务
特色服务

图拉扬分站

成都站、北京站
广州站、上海站
香港站

查看详情 >