机器学习驱动的法医感兴趣化合物色谱图、等离子图和红外光谱的数据融合

来源: Giorgio Felizzato et al 发布日期: 2025-02-21 访问量: 43

导读	在现场以尽可能高的精度获得法医分析的快速分析结果对于调查至关重要。虽然便携式传感器对于犯罪现场分析至关重要，但它们在灵敏度和特异性方面往往存在局限性，特别是由于环境因素。数据融合（DF）技术可以通过组合来自多个传感器的信息来提高准确性和可靠性。本研究利用离子迁移谱（IMS）和气相色谱石英增强光声光谱（GC-QEPAS）两种传感器的数据开发了不同的DF方法，旨在提高犯罪现场操作员的安全性和现场法医分析的准确性标签: 法医学、离子迁移谱、IMS、气相红外光谱、GC-QEPAS

Machine Learning-Driven Data Fusion of Chromatograms, Plasmagrams, and IR Spectra of Chemical Compounds of Forensic Interest

Giorgio Felizzato，Giuliano Iacobellis，Nicola Liberatore，Sandro Mengali，Martin Sabo，Patrizia Scandurra，Roberto Viola，Francesco Saverio Romolo*

Abstract

Achieving fast analytical results on-site with the highest possible accuracy in forensic analyses is crucial for investigations. While portable sensors are essential for crime scene analysis, they often face limitations in sensitivity and specificity, especially due to environmental factors. Data fusion (DF) techniques can enhance accuracy and reliability by combining information from multiple sensors. This study develops different DF approaches using data from two sensors: ion mobility spectrometry (IMS) and gas chromatography-quartz-enhanced photoacoustic spectroscopy (GC-QEPAS), aiming to improve the safety of crime scene operators and the accuracy of on-site forensic analysis. Two DF approaches were developed for acetone and DMMP: low-level (LLDF) and mid-level (MLDF), meanwhile a high-level (HLDF) approach was applied to TATP. LLDF concatenated preprocessed data matrices, while MLDF employed principal component analysis for feature extraction. LLDF and MLDF used one-class support vector machines (OC-SVM) for classification, while HLDF combined OC-SVM for IMS and SIMCA for GC-QEPAS. Sensor location within crime scenes was established using traditional measuring tape and laser distance meters, with a 1 m cutoff distance between sensors deemed appropriate for indoor crime scenes. LLDF achieved high accuracy but was sensitive to concentration variations, while MLDF enhanced the classification robustness. HLDF allowed for independent sensor use in real scenarios. All of the methods reached 100% accuracy for DMMP and acetone, and the MLDF approach was the fastest among the DF methods, demonstrating its potential for rapid applications. DF approaches can significantly enhance the safety and accuracy of forensic investigations, with future research planned to extend data sets and include more sensors.

在现场以尽可能高的精度获得法医分析的快速分析结果对于调查至关重要。虽然便携式传感器对于犯罪现场分析至关重要，但它们在灵敏度和特异性方面往往存在局限性，特别是由于环境因素。数据融合（DF）技术可以通过组合来自多个传感器的信息来提高准确性和可靠性。本研究利用离子迁移谱（IMS）和气相色谱石英增强光声光谱（GC-QEPAS）两种传感器的数据开发了不同的DF方法，旨在提高犯罪现场操作员的安全性和现场法医分析的准确性。针对丙酮和DMMP开发了两种DF方法：低级（LLDF）和中级（MLDF），同时将高级（HLDF）方法应用于TATP。LLDF将预处理后的数据矩阵连接起来，而MLDF采用主成分分析进行特征提取。LLDF和MLDF使用单类支持向量机（OC-SVM）进行分类，而HLDF将OC-SVM用于IMS，将SIMCA用于GC-QEPAS。犯罪现场内的传感器位置是使用传统的卷尺和激光测距仪确定的，传感器之间的截止距离为1米，被认为适合室内犯罪现场。LLDF实现了高精度，但对浓度变化很敏感，而MLDF增强了分类的鲁棒性。HLDF允许在真实场景中独立使用传感器。所有方法对DMMP和丙酮的准确率均达到100%，MLDF方法在DF方法中速度最快，证明了其快速应用的潜力。DF方法可以显著提高法医调查的安全性和准确性，未来的研究计划扩展数据集并包括更多的传感器。

···

For this purpose, several DF modules have been developed by employing two analytical sensors from the RISEN project (HORIZON2020, Grant Agreement No. 883116): an IMS sensor by MaSaTECH (Slovakia) (10) and a GC-QEPAS by Consorzio CREO (L’Aquila, Italy). (11,12) In the early stage of the study, the experimental data from the two sensors were studied and preprocessed to enhance reliability. Furthermore, various classification methodologies for each sensor were investigated to offer a comprehensive comparison between DF approaches and the performance of the sensors used alone. Following this step, the DF approach was focused on automating the classification task based on the sensors’ locations (their relative coordinates) within the crime scene. In this study, various machine learning (ML) methods were employed as tools for automating classification, including the one-class support vector machine (OC-SVM), the soft independent modeling of class analogy (SIMCA), and a combination of feature extraction using principal component analysis (PCA) with OC-SVM. In addition, the CPU time for each machine learning model was assessed to ensure the efficiency of our approaches in forensic scenarios. Indeed, in forensic investigations, the ability to quickly analyze data is just as important as accuracy, ensuring that timely decisions can be made based on the findings, especially in safety applications.

为此，利用RISEN项目（HORIZON2020，赠款协议号883116）的两个分析传感器开发了几个DF模块：MaSaTECH（斯洛伐克）的IMS传感器（10）和Consorzio CREO（意大利拉奎拉）的GC-QEPAS。（11,12）在研究的早期阶段，对两个传感器的实验数据进行了研究和预处理，以提高可靠性。此外，还研究了每种传感器的各种分类方法，以全面比较测向方法和单独使用的传感器的性能。在此步骤之后，DF方法侧重于根据传感器在犯罪现场内的位置（相对坐标）自动化分类任务。在这项研究中，各种机器学习（ML）方法被用作自动化分类的工具，包括单类支持向量机（OC-SVM）、类类比的软独立建模（SIMCA），以及使用主成分分析（PCA）和OC-SVM进行特征提取的组合。此外，还评估了每个机器学习模型的CPU时间，以确保我们的方法在取证场景中的效率。事实上，在法医调查中，快速分析数据的能力与准确性同样重要，确保根据调查结果及时做出决策，特别是在安全应用中。

The focus of the study was not the use of a large set of target analytes but rather the development of a new DF approach for real applications. The study has proven to be effective, opening new possibilities for improving safety and security during crime scene investigations.

这项研究的重点不是使用大量目标分析物，而是为实际应用开发一种新的DF方法。该研究已被证明是有效的，为改善犯罪现场调查期间的安全和安保开辟了新的可能性。

*本文内容非商业广告，仅供专业人士参考!