Identification of Bloodstains by Species Using Extreme Learning Machine and Hyperspectral Imaging Technology

Area of Science:

• Forensic Science
• Spectroscopy
• Machine Learning

Background:

• Accurate bloodstain identification is crucial for criminal investigations.
• Distinguishing between human and animal bloodstains can be challenging.
• Current methods may be destructive or time-consuming.

Purpose of the Study:

• To develop a rapid, non-destructive method for identifying bloodstain species.
• To evaluate the effectiveness of hyperspectral imaging combined with machine learning algorithms.
• To compare the performance of extreme learning machine (ELM) with support vector machine (SVM) and random forest (RF) algorithms.

Main Methods:

• Acquired spectral data of human and animal bloodstains using hyperspectral imaging.
• Developed and trained classification models using the extreme learning machine (ELM) algorithm.
• Compared ELM performance against support vector machine (SVM) and random forest (RF) algorithms.

Main Results:

• The extreme learning machine (ELM) algorithm demonstrated superior performance.
• ELM achieved the highest precision, sensitivity, specificity, and F1 score.
• Hyperspectral technology combined with ELM provided accurate bloodstain species identification.

Conclusions:

• Hyperspectral imaging and ELM offer a rapid, non-destructive, and accurate solution for bloodstain identification.
• This approach provides a valuable new technical reference for forensic analysis.
• The findings support the use of advanced spectral and machine learning techniques in criminal investigations.