激光诱导击穿光谱
稳健性(进化)
计算机科学
光谱学
元素分析
激光烧蚀
模式识别(心理学)
生物系统
材料科学
人工智能
数据挖掘
机器学习
激光器
化学
光学
物理
生物
量子力学
基因
有机化学
生物化学
作者
J. Cárdenas-Escudero,David Galán-Madruga,Jorge O. Cáceres
标识
DOI:10.1177/00037028241277897
摘要
This article provides a detailed discussion of the evidence available to date on the application of laser-induced breakdown spectroscopy (LIBS) and supervised classification methods for the individual reassignment of commingled bone remains. Specialized bone chemistry studies have demonstrated the suitability of bone elemental composition as a distinct individual identifier. Given the widely documented ability of the LIBS technique to provide elemental emission spectra that are considered elemental fingerprints of the samples analyzed, the analytical potential of this technique has been assessed for the investigation of the contexts of commingled bone remains for their individual reassignment. The LIBS bone analysis consists of the direct ablation of micrometric portions of bone samples, either on their surface or within their internal structure. To produce reliable, accurate, and robust bone classifications, however, the available evidence suggests that LIBS spectral information must be processed by appropriate methods. When comparing the performance of seven different supervised classification methods using spectrochemical LIBS data for individual reassociation, those employing artificial intelligence-based algorithms produce analytically conclusive results, concretely individual reassociations with 100% accuracy, sensitivity, and robustness. Compared to LIBS, other techniques used for the purpose of interest exhibit limited performance in terms of robustness, sensitivity, and accuracy, as well as variations in these results depending on the type of bones used in the classification. The available literature supports the suitability of the LIBS technique for reliable individual reassociation of bone remains in a fast, simple, and cost-effective manner without the need for complicated sample processing.
科研通智能强力驱动
Strongly Powered by AbleSci AI