| Autor: |
Aleksandr D. Afanasyev, Ily’a O. Prichko |
| Rok vydání: |
2021 |
| Předmět: |
|
| Zdroj: |
Studies in Systems, Decision and Control ISBN: 9783030564322 |
| DOI: |
10.1007/978-3-030-56433-9_12 |
| Popis: |
The purpose of the paper is to demonstrate the applicability of artificial intelligence methods for the identification of relevant forensic information from records of video surveillance systems and the opportunity to classify this information by the level of forensic relevance from potentially significant to less important events. Methodology: The paper deals with common problems related to the work of a criminal investigator, provides their analysis, and offers solutions based on modern AI developments. The solutions are justified by the new capabilities of mathematical simulation and numerical methods and specially designed programs. Findings: The paper offers a comprehensive solution for the identification of relevant forensic information, which allows achieving high efficiency against conventional approaches as well as reducing man-caused errors. It has been remarked that the most usable form of an efficient video review tool is a comprehensive solution involving the following processing stages: (1) Reducing non-informative events to a minimum and using video frame space in an efficient way; (2) Classification of objects and search by required criteria; (3) Identification and re-identification of people between all sources of video streams. This three-stage solution will help the investigator to focus only on the events that are relevant to the investigation. Besides, there will be new opportunities against conventional methods. The paper discusses some constraints that may serve as a guide for future research. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
