IDENEO

IDENEO

Identification of unusual behavior
of people in video streams.

PROJECT GENERAL INFORMATION

Contract Number

Project Code

PN-III-P2-2.1-SOL-2021-0024

Project Name

Identification of unusual behavior of people in video streams.

Acronym

IDENEO

The IDENEO project "Identifying the unusual behavior of people in video streams" aims to develop AI algorithms capable of detecting suspicious behavior of people and crowds in order to combat violent or terrorist actions in certain circumstances. The proposed solution will allow the authorities to detect potential dangers in various and dynamic scenarios, such as protests or security of monitored areas of interest, and to understand the entire operational scenario, by combining information from cameras, analyzing images and displaying them on the GIS map.

ACTIVITIES

Stage 1

Investigation of State-of-the-art technologies. Identification of the limits, the constraints and the improvement possibilities of the technology that will be adopted in future implementations correlated with the specific requirements.

hologram

Stage 2

Design, Development and Implementation of innovative algorithms and software modules that will be part of the final solution.

programming

Stage 3

Developing the prototype. Testing and validation.

OBJECTIVES

Detection and analysis of events.

Behavior detection and analysis.

  • Detection of persons who do not respect social distance (with a threshold selectable by the user).

  • Detection of the groups (clusters) of people from the same neighborhood and the attributes related to them (number, density, adults / children).

  • Estimation of the direction and speed of movement of people.

  • Estimation of the direction and speed of movement of the group.

  • Estimation of the gradient of people density (increasing or decreasing density).

  • Detection of long term stationary phases.

  • Identify objects of interest that appear in the monitored area (which can generate errors: False Positive (a swing, a coat hanger, etc.) or False Negative (undetected people).

  • Estimation of the distance between groups of people.

  • Geolocation of groups.

REZULTATE​

Stage 1

Prima etapă a proiectului a avut în vedere studiul și evaluarea tehnologiilor State-of-the-Art pentru dezvoltarea algoritmilor de detecție precum și identificarea limitărilor/constrângerilor de natura algoritmica/tehnica și a posibilităților de îmbunătățire a tehnologiei ce va fi adoptată în implementările viitoare corelată cu cerințele specifice. În cadrul acestei etape au fost prevăzute 7 activități care au fost realizate cu informarea și în parteneriat cu beneficiarul More information

Stage 2

Etapa II are ca obiective dezvoltarea algoritmilor AI care corespund scenariilor din proiect, la proiectarea și dezvoltarea principalelor componente software ale sistemului și la integrarea modulelor prototip într-un produs software cu nivel de maturitate TRL5. More information

Stage 3

Etapa III a avut ca principale obiective dezvoltarea unei soluții software (TRL7) pentru identificarea comportamentului neobișnuit al persoanelor în fluxurile video in timp real, testarea și validarea acesteia în condiții reale precum si elaborarea unei strategii de exploatare a rezultatelor. Activitățile derulate au inclus testarea, validarea, optimizarea modulelor dezvoltate și integrarea acestora în cadrul platformei K-Vision. More information

CONSORTIUM

THE TEAM OF "OVIDIUS" UNIVERSITY OF CONSTANTA

Dragoș SBURLAN
Dragoș SBURLAN
UEF-ID: U-1700-030V-4710 <br>Project Director
Dragoș Sburlan is Associate Professor of Computer Science at the Ovidius University of Constanta. His main interests regard the development of new computing paradigms and algorithms in the field of theoretical computer science.
Elena PELICAN
Elena PELICAN
UEF-ID: U-1700-035Q-6523<br>Researcher
Elena Pelican is Associate Professor at the Faculty of Mathematics and Informatics, Ovidius University, Constanta. Her actual research interest focuses upon deep learning for computer vision and natural language processing problems, pattern recognition, and more broadly, image processing and machine learning.
Dorin-Mircea POPOVICI
Dorin-Mircea POPOVICI
UEF-ID: U-1700-039W-6468<br>Researcher
Dorin-Mircea POPOVICI is a university professor. PhD at the Faculty of Mathematics and Informatics of Ovidius University of Constanta, coordinator of the Virtual and Augmented Reality Research Laboratory (CeRVA) - http://cerva.ro. PhD supervisor in the field of Computers and Information Technology has as fields of interest Virtual, Augmented and Mixed Reality Environments for Education, Training and Cultural Heritage, Multimodal Interfaces and Multi-agent Systems.
Gabriel PRODAN
Gabriel PRODAN
UEF-ID: U-1800-048D-1500<br>Researcher
Alexandru BOBE
Alexandru BOBE
UEF-ID: U-1700-030D-8973<br>Researcher
George PRODAN
George PRODAN
George Prodan received the B.S. degree in Computer Science from the Ovidius University of Constanta. His main interests are image and language processing for Big Data and their general applications.
Sebastian RĂDUCAN
Sebastian RĂDUCAN
Student of Ovidius University of Constanta, Răducan Sebastian's main interests lie in Computer Vision and Face Recognition.

SOFTRUST VISION ANALYTICS TEAM

Marian GHENESCU
Marian GHENESCU
Roxana-Elena MIHĂESCU
Roxana-Elena MIHĂESCU
Șerban CARATĂ
Șerban CARATĂ

TEAM OF THE POLYTECHNIC UNIVERSITY OF BUCHAREST

Bogdan IONESCU
Bogdan IONESCU
Liviu-Daniel ȘTEFAN
Liviu-Daniel ȘTEFAN
 Mihai-Gabriel CONSTANTIN
Mihai-Gabriel CONSTANTIN
Mihai DOGARIU
Mihai DOGARIU

CONTACT

Universitatea "Ovidius" din Constanța