FORENSOR

European Commission
Funding Organization:	European Commission
Funding Programme:	Horizon H2020
Funding Instrument:	Innovation Action
Start Date:	September 2, 2015
Duration:	36 months
Total Budget:	4,043,546 EUR
ITI Budget:	611,500 EUR

FOREnsic evidence gathering autonomous seNSOR (2015: H2020 IA)

Law enforcement Agencies (LEAs) are even today using conventional, man-power based techniques such as interviews and searches to gather forensic evidence. As sophistication of both criminals and their crimes are increasing, investigators must improve the means available for gathering a body of compelling evidence of a suspect’s involvement in a crime. Concealed surveillance devices have been instrumental in this direction, providing irrefutable evidence that can play an important part in bringing criminals to justice. However, current video surveillance systems are usually bulky and complicated, and rely on complex, expensive infrastructure to supply power, bandwidth and illumination.

Over 36 months, FORENSOR -FOREnsic evidence gathering autonomous seNSOR- will develop and validate a novel, ultra-low-power, miniaturised, low-cost, wireless, autonomous sensor (“FORENSOR”) for evidence gathering, able to operate for up to two months without infrastructure. FORENSOR will be manageable remotely, will preserve the availability and the integrity of the evidence collected, and comply with all legal and ethical standards, in particular those related to privacy and personal data protection. Secure and intelligent communications let such sensors join their forces towards robust evidence management and real time monitoring and control operations. The combination of built-in intelligence with ultra-low power consumption will make this device a true breakthrough for combating crime.

FORENSOR will combine:

a) Built-in intelligenceto allow acquisition of suspected criminal activity only.
b) Ultra-low power operationenabling operation for several weeks, even in infrastructure-less environments.
c) Operational adaptivityallowing the sensor to adjust to different operational requirements and use cases. The FORENSOR will work in both remote, low-activity environments, working for months on a single battery w/wo power harvesting, and in crowded environments where the sensor will be able to operate autonomously for several weeks.
d) Remote managementfor receiving crime alerts, monitoring the sensor’s operational status, managing the operating profiles and the recorded data.
e) Small form factorfor easy deployment and concealment in environments.
f) Data integrityto ensure that the chain custody remains intact and the integrity of the evidence will be kept until the data is presented in a court of law.
g) Robustnessto ensure the availability and the integrity of the evidence recorded.
h) Secure communicationsfor smooth collaboration among sensors, reliable evidence management and real time operation between the sensors and the monitoring station.
i) Low costto make viable the deployment of swarms of sensors.

Publications:

A. Doumanoglou, N. Vretos, P. Daras, "Frequency–Based Slow Feature Analysis" , Neurocomputing, 368, 34-50, 2019. DOI: https://doi.org/10.1016/j.neucom.2019.08.067

G. Zoumpourlis, A. Doumanoglou, N. Vretos, P. Daras, "Non-linear Convolution Filters for CNN-based Learning" , IEEE International Conference on Computer Vision (ICCV 2017), Venice, Italy, October 22-29 2017

A. Doumanoglou, N. Vretos, P. Daras, "Action Recognition From Videos using Sparse Trajectories" ,7th International Conference on Imaging for Crime Detection and Prevention (ICDP-16), Madrid, 23-25 November, 2016

Contact:
Dr. Petros Daras

Visit the website:
forensor-project.eu

News Releases!

FORENSOR - Results in Brief article on CORDIS

July 10, 2019

Set-and-forget surveillance cameras wake themselves up and call for help Despite their advantages, digital security cameras are inconvenient. A new surveillance model reports on crime by itself. Digital surveillance cameras have come a long way in a short time. They now combine very-high–quality images with autonomous operation and convenient sizes. Therefore, such devices are widely used in law enforcement. Nevertheless, video surveillance remains costly, power hungry and complicated. Current systems are used mainly for recording and have very little capability for real-time scene interpretation. The EU-funded FORENSOR project has developed a new sensor system that addresses these problems. The sensor is small and concealable; its low-power requirements free it from electrical infrastructure. The device can operate for months. Hence, the sensor is also capable of a degree of independent operation. You can read the full article here.

Continue Reading...
FORENSOR: Action Recognition

May 09, 2017

Action Recognition via Motion Analysis In this method, the first step of performing real-time object tracking is based on background subtraction. Initially, the algorithm grabs an image of the background, assuming that the frame is empty of subjects. Subsequently, while capturing each new frame, the background is subtracted by the newly acquired image and the resulted outcome is thresholded to form a foreground binary mask. This mask is further enhanced by the morphological operation of dilation and finally a connected component analysis is utilized to obtained the final foreground label map. This label map is used as a basis for object tracking. In this method, action recognition classification is performed by tracking the trajectory of the foreground bounding box centroid. This method classifies the action of the foreground object between walking and standing. [video width="854" height="480" mp4="http://vcl.iti.gr/vclwordpress/wp-content/uploads/2017/05/action_recognition_motion_analysis.mp4"][/video] Action Recognition via Motion Tracking This method constitutes a novel, low-complexity, method for action recognition from videos. Initially, a 3D Sobel filter is applied to the video volume resulting into a binary image with non-zero pixels in areas of motion. The non-zero valued pixels are spatially clustered using k-means and the most dominant centers of video motion are extracted. The centers are then tracked forming sparse trajectories, whose properties are later used to create a new feature type, namely the Histogram of Oriented Trajectories (HOT), describing the video. Feature vectors are finally passed to an AdaBoost classifier for classification. The proposed method reports competitive results in KTH and MuHAVi datasets, while remaining low in complexity and thus being suitable to be used in surveillance systems requiring low processing power. In the provided video, trajectory extraction and the corresponding classification result is presented. This work, (“Action Recognition From Videos using Sparse Trajectories”) has been published in the 7th International Conference on Imaging for Crime Detection and Prevention (ICDP-16), Madrid, 23-25 November, 2016. [video width="320" height="240" mp4="http://vcl.iti.gr/vclwordpress/wp-content/uploads/2017/05/action_recognition_motion_tracking.mp4"][/video]

Continue Reading...
VCL participates in two new H2020 projects: FORENSOR & TRILLION

September 16, 2015

Two new EU funded H2020 projects have just launched. VCL coordinates the project: FORENSOR — FOREnsic evidence gathering autonomous seNSOR (2015: H2020 IA) TRILLION — Trusted, Citizen, LEA collaboration over social Networks (2015: H2020 RIA)

Continue Reading...