학술논문


A first step towards a machine learning-based framework for bloodstain classification in forensic science
Document Type
Article
Source
In Forensic Science International December 2024 365
Subject
Language
ISSN
0379-0738
Abstract
Highlights •A preprocessing of bloodstain dataset. The background of the photo of bloodstain was removed using the Otsu method technique to better highlight the characteristics of the bloodstains. (see Section 4).•A feature analysis for bloodstains with five different characteristics related to bloodstains such as the bloodstain’s angle, shape, major and minor axis lengths, and the overall bloodstain distribution. (see Section 5).•An ML-based classification framework with three ML models such as the CNN, Decision Tree, and Random Forest. (see Section 6).