• Unsupervised machine learning...

*Result*: Unsupervised machine learning for the detection and interpretation of key features in drip patterns.

Title:
Unsupervised machine learning for the detection and interpretation of key features in drip patterns.
Authors:
Pachong SM; Faculty of Business and Information Technology, Ontario Tech University, 2000 Simcoe St N, Oshawa, Canada; Faculty of Science, Ontario Tech University, 2000 Simcoe St N, Oshawa, Ontario L1G 0C5, Canada., Alavi A; Faculty of Business and Information Technology, Ontario Tech University, 2000 Simcoe St N, Oshawa, Canada; Faculty of Science, Ontario Tech University, 2000 Simcoe St N, Oshawa, Ontario L1G 0C5, Canada., Kannan S; Faculty of Science, Ontario Tech University, 2000 Simcoe St N, Oshawa, Ontario L1G 0C5, Canada., Stotesbury T; Faculty of Science, Ontario Tech University, 2000 Simcoe St N, Oshawa, Ontario L1G 0C5, Canada., Lewis PR; Faculty of Business and Information Technology, Ontario Tech University, 2000 Simcoe St N, Oshawa, Canada. Electronic address: peter.lewis@ontariotechu.ca.
Source:
Forensic science international [Forensic Sci Int] 2026 Feb; Vol. 378, pp. 112669. Date of Electronic Publication: 2025 Sep 25.
Publication Type:
Journal Article
Language:
English
Journal Info:
Publisher: Elsevier Science Ireland Country of Publication: Ireland NLM ID: 7902034 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1872-6283 (Electronic) Linking ISSN: 03790738 NLM ISO Abbreviation: Forensic Sci Int Subsets: MEDLINE
Imprint Name(s):
Publication: Limerick : Elsevier Science Ireland
Original Publication: Lausanne, Elsevier Sequoia.
MeSH Terms:
Blood Stains* , Unsupervised Machine Learning* , Image Processing, Computer-Assisted* , Machine Learning*, Humans
Contributed Indexing:
Keywords: Bloodstain pattern analysis; Machine learning; Unsupervised learning
Entry Date(s):
Date Created: 20251001 Date Completed: 20251204 Latest Revision: 20251204
Update Code:
20260130
DOI:
10.1016/j.forsciint.2025.112669
PMID:
41033067
Database:
MEDLINE

*Further Information*

*Bloodstain pattern analysis (BPA) is increasingly shifting towards more objective methodologies for pattern classification. This transition can involve image-processing techniques that extract observable bloodstain features as data for pattern classification. This paper explores how unsupervised machine learning (ML)-based frameworks can be designed to identify observable features in bloodstain patterns, starting with a basic drip pattern. A total of 398 laboratory-generated drip patterns were analyzed, spanning dripping heights between 25 and 100 cm and droplet counts ranging from 1 to 10. The extracted observable features incorporated key bloodstain properties commonly used in forensic analysis, such as size and shape, hence aligning with previously reported qualitative properties and existing bloodstain taxonomies. To assess feature importance, SHAP (SHapley Additive exPlanations) analysis was applied, ranking features by their contributive power to the model's predictions. The results revealed that the circularity, the mean intensity, and the area of the parent stain were the three most significant features for distinguishing drip patterns with contribution power of 60 %, 28 %, and 28 %, respectively, when excluding the dripping height and the number of droplets from the model. This unsupervised ML-driven approach demonstrates strong potential for establishing feature criteria for image-processing based bloodstain pattern classification methods.
(Copyright © 2025. Published by Elsevier B.V.)*

*Declaration of Competing Interest The authors hereby declare having no known competing financial or personal interests that could have influenced the work in this paper.*