Clustering and Unsupervised Classification in Forensics
From Theory to Practice
Nowadays, crime investigators collect an ever increasing amount of potential digital evidence from suspects, continuously increasing the need for techniques of digital forensics. Often, digital evidence will be in the form of mostly unstructured and unlabeled data and
seemingly uncorrelated information. Manually sorting out and understanding this type of data constitutes a considerable challenge, sometimes even a psychological burden, or at least a prohibitively time consuming activity. Therefore, forensic research should explore
and leverage the capabilities of cluster algorithms and unsupervised machine learning to-wards creating robust and autonomous analysis tools for criminal investigators faced with this situation. This report presents a first comprehensive study from theory to practice on
the specific case of video forensics.
JUNKLEWITZ Henrik;
FERRARA Pasquale;
BESLAY Laurent;
2021-03-11
Publications Office of the European Union
JRC119038
978-92-76-23872-0 (online),
1831-9424 (online),
EUR 30419 EN,
OP KJ-NA-30419-EN-N (online),
https://publications.jrc.ec.europa.eu/repository/handle/JRC119038,
10.2760/308387 (online),
Additional supporting files
| File name | Description | File type | |
|---|
