SURV-AI-LLANCE
  • Home
  • Our Goal
  • Our Approach
  • Team
  • Conclusions
  • Closing event
  • Home
  • Our Goal
  • Our Approach
  • Team
  • Conclusions
  • Closing event

Alert officials
​in real-time using behavioral insights

With support from
Picture
 

Our Goal: Truly Smart Surveillance

Current urban surveillance systems are monitored intensively and round-the-clock by qualified personnel. But the large volume of generated video data is near impossible to process in real-time by human observation alone. Nevertheless, responding swiftly to situations involving violence is essential to provide safe environments that protect our citizens.

The Surv-AI-llance project aims to gather reliable and actionable insights into human behavior by developing advanced machine-learning algorithms based on data from cameras and radar devices. These algorithms will be capable of interpreting scenes involving violence and aggression, enabling real-time alerts and rapid response from safety and security officers while guaranteeing the privacy of the people involved.

The overarching objective of the Surv-AI-llance consortium is to build a reliable, privacy-friendly video analytics pipeline that accurately interprets surveillance scenes under surveillance and rapidly alerts officials. 
​
Picture
 

Our Approach

Picture

Combining camera and radar technology

Collect data from videosurveillance and radar technology.

Important
We use skeleton overlays for this research. All data is anonymous.
Picture

Teaching software to recognise human behaviour

A team of dedicated people analyse all gathered data and classify it under the right behaviour. After a while, the Surv-AI-llance software itself will be able to identify human behaviour.
Picture

Building triggers which alert officials instantly

The triggers in the Surv-AI-llance software instantly alert officials whenever certain behaviour is detected.​ This dissolves the need to monitor videosurveillance round-the-clock. 
 

Team

​The imec.icon project, Surv-AI-llance, is a research project bringing together
​academic researchers, imec - ETRO - VUB  and  imec - SENDA, and six industry partners:
Picture
www.imec.be
Picture
www.etrovub.be
Picture
www.vub.be
Picture

Autimatic

Autimatic will establish semi-automated ways to process massive amounts of video data into training data. This team consists of people within the Autism spectrum, who named their services accordingly: Autism Augmented Artificial Intelligence.
Picture

​Atomic BITS

Atomic BITS will use Automatic's training datasets, combined with the gathered radar data, to create and continuously improve the deep learning algorithms that perform the behavior analysis. 
Picture

The Safe Group

The Safe Group will integrate Atomic BITS' software into the existing video management system. This creates an innovative workflow: the human behavior detection system will lead to very efficient processes for surveillance personnel to alert, categorize, dispatch and follow-up on crime in a timely manner.
Picture

Local Police Meetjesland Centrum

End-user who will be testing in the field.
Picture

Local Police Geel-Laakdal-Meerhout

End-user who will be testing in the field.
Picture

City of Geel

End-user who will be testing in the field.
​Surv-AI-llance is co-financed by imec and receives financial support from Flanders Innovation & Entrepreneurship (project nr. HBC.2020.3106).
 

Conclusions

The closing session "Prediction Human Movements in Public Space." will be hosted Friday 23 February 2024. During this event the research findings will be discussed. Following academic papers will be highlighted. The industry partners will discuss their valorization options as well as their plan-of-action for the next 12 months.

​Publication list:
  • Y. Chen, N. Deligiannis, "Locally accumulated Adam for Distributed training with Sparse Updates," in IEEE International Conference on Image Processing (ICIP), 2023.
  • A. Stergiou, N. Deligiannis, "Leaping Into Memories: Space-Time Deep Feature Synthesis," in IEEE International Conference on Computer Vision (ICCV), 2023.
  • X. Zhang, B. Joukovsky, N. Deligiannis, "Quantitative Evaluation of Video Explainability Methods via Anomaly Localization," in European Signal Processing Conference (EUSIPCO), 2023.
  • B. De Weerdt, Y. C. Eldar, N. Deligiannis, "Designing Transformer networks for sparse recovery of sequential data using deep unfolding," in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2023.
  • A. Stergiou, B. De Weerdt, N. Deligiannis, "Holistic Representation Learning for Multitask Trajectory Anomaly Detection," in IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2024. 
  • S. Hamed Javadi, Hichem Sahli, and André Bourdoux, “Rayleigh-based segmentation of ISAR images”, Applied Optics Vol. 62, Issue 17, pp. F1-F7 (2023).
  • Hamed Javadi, Ruoyu Feng, André Bourdoux, and Hichem Sahli. "Multi-target tracking pipeline for MIMO-FMCW radars based on modified GM-PHD." In 2023 31st European Signal Processing Conference (EUSIPCO), pp. 1818-1822, IEEE (2023).
​
 
 

Join the closing event

​The closing session is open to the public, but registration is required.
​LOCATION:
Corda Campus, building 1 (Corda Conference: floor 1)
Kempische Steenweg 293/16,
3500 Hasselt

PARKING:
Free entrance for visitors


PUBLIC TRANSPORT:
Easy to reach from railway station Kiewit
DATE:
Friday 23 February 2024

​
PROGRAMME:
14:00 / Welcome with industry & research experts
14:15   / Opening presentation by the industry
​14:45  / Presenting the research papers
16:00 / Demonstrations + Q&A
17:00 / Closing
Submit

Home

Our Goal

Our Approach

Team

Contact

Copyright © 2022
Photo from Mary Vican