Powell Rees, van Bronswijk Wilhelm, Coumbaros John
Forensic Chemistry Section, ChemCentre, Level 2, South Wing, Building 500, Resources and Chemistry Precinct, Cnr Manning Rd and Townsing Dr, Bentley, Western Australia 6102, Australia.
Department of Chemistry, Curtin University, GPO Box U1987, Perth, Western Australia 6845, Australia.
Forensic Sci Int. 2018 Jun;287:54-62. doi: 10.1016/j.forsciint.2018.03.025. Epub 2018 Mar 20.
In numerous major crime cases undertaken at our laboratory the recovery of large numbers of fibres (>1000), particularly in cases with no known source, presented several challenges. These included the inability to effectively manage the data (i.e. combination of MSP spectra, microscopic characteristics, composition, etc.) and perform comparisons in an efficient manner. To address these challenges, and in response to a growing need for performing fibre comparisons, we developed a database of textile fibre microspectrophotometric (MSP) spectra. The database, designed to compare MSP spectra using a modified Pearson method of correlation, currently contains over 20,000 normalised and first derivative spectra of casework, validation and reference textile fibres. A comparison strategy for cases with a large number of questioned samples was devised, involving identification of critical fibres in the casework data set, development of preliminary fibre groups classified according to their corresponding/similar MSP spectra, and verification of the preliminary groups via brightfield and fluorescence comparison microscopy. The database has successfully been utilised for proficiency trials and casework with small questioned fibre sets. Furthermore, in a case involving a larger dataset (>4000 "unknown" fibres) the database assisted in the efficient classification of 156 distinct groups of interest, highlighting its utility in providing investigative leads for the identification of potential sources of the recovered fibres.
在我们实验室承办的众多重大犯罪案件中，大量纤维（>1000）的回收，尤其是在来源不明的案件中，带来了诸多挑战。这些挑战包括无法有效管理数据（即质谱图、微观特征、成分等的组合）以及无法高效地进行比对。为应对这些挑战，并响应日益增长的纤维比对需求，我们开发了一个纺织纤维显微分光光度法（MSP）光谱数据库。该数据库旨在使用改进的皮尔逊相关方法比较MSP光谱，目前包含超过20,000条来自实际案件、验证和参考纺织纤维的归一化和一阶导数光谱。针对有大量可疑样本的案件设计了一种比对策略，包括识别实际案件数据集中的关键纤维、根据其相应/相似的MSP光谱开发初步纤维组，以及通过明场和荧光比较显微镜对初步纤维组进行验证。该数据库已成功用于能力验证试验和涉及少量可疑纤维组的实际案件。此外，在一个涉及更大数据集（>4000根“未知”纤维）的案件中，该数据库协助对156个不同的感兴趣组进行了有效分类，凸显了其在为识别回收纤维潜在来源提供调查线索方面的实用性。
