Rapid Non-Destructive Characterization of Fingernails by Attenuated Total Reflection–Fourier Transform Infrared Spectroscopy (ATR–FTIR) and Deep Learning

Abstract

In the realm of forensic anthropology, fingernails serve as invaluable repositories of biological information due to their numerous advantages. To facilitate precise identification in terms of gender and geographic origin, we conducted a study employing fingernails from various regions of the Chinese mainland (specifically, Gansu, Guilin, Hebei, Liaoning, and Jiangsu). Our comprehensive approach amalgamates attenuated total reflection–Fourier transform infrared spectroscopy (ATR–FTIR) with decision tree (DT), support vector machine (SVM), and random forest (RF). This approach boasts the merits of speed and nondestructiveness. The C5.0-based DT model and the SVM models based on both radial basis function (RBF) and linear kernels achieve a favorable 100% accuracy in distinguishing between genders and geographic origins. Furthermore, the optimized RBF model exhibits superior performance in regional differentiation, particularly for male samples. In conclusion, the utilization of ATR-FTIR for the swift differentiation of fingernails based on gender and geographic origin is indeed feasible. This paves the way for the meaningful application of machine learning algorithms in the realm of forensic anthropology, offering valuable insights and prospects for future research.

Keywords:

• Attenuated total reflection–Fourier transform infrared spectroscopy (ATR–FTIR
• decision tree
• fingernails
• random forest
• support vector machine

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the General Program of Hunan Provincial Natural Science Foundation of China (2023JJ30221).