Non-invasive prediction of bloodstain age using the principal component and a back propagation artificial neural network. (18th August 2017)

• Record Type: Journal Article
• Authors: Sun, Huimin; Meng, Yaoyong; Zhang, Pingli; Li, Yajing; Li, Nan; Li, Caiyun; Guo, Zhiyou
• Is Part Of: Laser physics letters. Volume 14:Number 9(2017:Sep.)
• Publication Date: 2017-08-18
• Subjects: whole blood -- Raman spectroscopy -- PCA -- BP-ANN -- forensics; Lasers -- Periodicals; Physics -- Periodicals; 621.366
• DOI: 10.1088/1612-202X/aa7c48
• Languages: English
• ISSNs: 1612-2011

Abstract

Abstract: The age determination of bloodstains is an important and immediate challenge for forensic science. No reliable methods are currently available for estimating the age of bloodstains. Here we report a method for determining the age of bloodstains at different storage temperatures. Bloodstains were stored at 37 °C, 25 °C, 4 °C, and −20 °C for 80 d. Bloodstains were measured using Raman spectroscopy at various time points. The principal component and a back propagation artificial neural network model were then established for estimating the age of the bloodstains. The results were ideal; the square of correlation coefficient was up to 0.99 ( R 2 > 0.99) and the root mean square error of the prediction at lowest reached 55.9829 h. This method is real-time, non-invasive, non-destructive and highly efficiency. It may well prove that Raman spectroscopy is a promising tool for the estimation of the age of bloodstains.