CORDIS - Forschungsergebnisse der EU

Petty cRiminality diminution through sEarch and Analysis in multi-source video Capturing and archiving plaTform

Final Report Summary - P-REACT (Petty cRiminality diminution through sEarch and Analysis in multi-source video Capturing and archiving plaTform)

Executive Summary:
The P-REACT project has met its originally planned objectives, the consortium is satisfied with the work done and with the results achieved. These results, when presented to the stakeholders in the two final trials have been well received and raise promising exploitation opportunities.
By definition, all plans are just a way to organise future work. Reality however is complex and full of unforeseen events that impact on the initial plans. Deviations therefore are natural and have to be managed in a way that does not affect the essence of the objectives and goals of the project. In the P-REACT project there have been multiple deviations from the original plan, some were decided by the consortium, and some others were unpredicted and required a reactive response to mitigate its negative impact or to maximise the positive effect.
By the end of the project, the P-REACT solution, which consists of an advanced prototype that has been demonstrated with simulated events in operational environments, meets all the initial goals that were set out at the beginning of the project.
The P-REACT solution is a cost-effective solution for the protection of small businesses and transit infrastructure in an urban environment from petty crimes. It provides situational awareness and incident classification functionality to end-users based on the utilisation of privacy-conserving data from local embedded sensors. It includes a novel visual analytics framework for addressing early diagnosis of incidents and enriched offline multimodal visualisation mechanisms to strengthen decision-making mechanisms. It is based on a novel cloud-based service Video Surveillance as a Service (VSaaS). The solution consists of a novel low-cost embedded framework with local intelligence and real-time alert categorisation. It exploits the use of semantic-technologies with evidence from video archive sources in order to forecast, prevent and detect future petty crimes. It has been designed, developed, and demonstrated in compliance with the legal, ethical, and privacy regulations and recommendations (including the external guidance and audit of three experts). The solution has been demonstrated in two final public trials.
Beyond meeting all these initial goals, the P-REACT solution has been developed so that it is interoperable and easy to integrate with other systems or existing infrastructures (there is no vendor lock-in). It is modular and scalable in a way that additional analysis modules from existing or new types of sensors can easily be included. It is dynamically reconfigurable because it includes business rules which allow setting the default behaviours of the system to be modified based on the context and the events that are detected.
The P-REACT solution is not however a mature product which can directly be launched to market or put in a production environment. Tests and demonstrations conducted in controlled environments are promising but further evaluation of the system in real environments is needed to fully determine the level of reliability of its petty crimes detection capabilities. At the time of writing this report, the consortium is analysing the possibility of running one or two pilots of the system in real scenarios (more than a dozen companies have shown interest in offering their premises for these pilots and the consortium is analysing the effort, costs, pros and cons of undertaking this pilots after the finalisation of the project.

Project Context and Objectives:
In this section we briefly present the main objective and the scientific and technical objectives of the project as they were established during the proposal preparation phase of the project. In the following section we will describe the main results produced during the project and how they map to the objectives presented in this section.
1.1.1 Initial context
Petty crime, also known as volume crime, is crime that occurs frequently and affects many people. If law enforcement agencies (LEA) are to have an impact on petty crime, then effective deterrent and investigation should form a key part of a crime reduction strategy.
Petty crime imposes on community daily life . Unsurprisingly it has been most frequent in urban and busy transit areas. Police attribute the steep increase to variety of socioeconomic factors (low wealth level of local community, increased unemployment rate, youth gangs, drug addicts, illegal immigrants etc.).
Citizens and communities perceive petty crime to be a key issue. It is seen as a national security issue by 13% of Europeans. On a daily basis, the media report on crime, highlighted it as a key issue and reinforcing its negativity to the public. The result is that even a low-intensity has a negative effect on the citizen psychology.
Currently the most common solution for protection of small businesses, communities and citizens are video surveillance systems (CCTV). In the last decade there has been a proliferation of video surveillance installations, particularly in town centres, shops, shopping centres, banks, building societies, parking facilities, schools, colleges, hospitals, transport facilities, industrial estates, business centres, football grounds etc. However, it is evident that the installed base of surveillance systems is comprised of diverse and non-interoperable systems which are not ‘smart’, capable of providing alerts nor easily accessible by law enforcement. Existing systems installed in open areas or infrastructure store data in non-interoperable archiving systems thus making their exploitation by LEA nearly impossible.
A key factor for the reduction of petty crimes is the fostering of relationships between the local community and law enforcement agencies. It is important that incidents are reported to the police and that the local community recognises the importance of working with the authorities on a partnership basis for the reduction of petty crime, constituting a cycle of trust.
The current status of the video surveillance systems (CCTV) for petty crimes protection is summarised in the following key points:
1. High installation costs of surveillance solutions for small business etc.
2. No connectivity with law enforcement agencies.
3. Lack of early warning mechanisms.
4. Lack of unified archiving system integrating multi-source video.
5. Legacy systems are not supported by video analytics applications in terms of integration.
6. There is a lack of standards that address tagging of video archive content with appropriate metadata.
7. The validity of video archive content presented as evidence to authorities must respect a chain of custody.
8. Currently, there is still a need to physically transport video surveillance footage while the operator manually analyses the video.
9. Currently there is no system that manages automated search and correlation requiring significant manual effort to trawl through video archives when requested to present video footage to law enforcement agencies and judicial authorities.

1.1.2 Main objective of the P-REACT project
The Description of Work of the P-REACT proposal describes the project’s main objective as: “The P-REACT project proposes the creation of a video capturing and archiving network/platform that enables the protection of small businesses and transit locations in an urban environment from petty crimes. It will be based on low-cost components and will build on the capabilities offered by both established and emerging technologies, from Broadband internet to Cloud computing. The basic idea is to install low-cost components in the small business premises and transit areas, that are networked directly to, one or more, Cloud-based, Data Centres, where their activity is constantly monitored and recorded. The quality of the captured and recorded video will be based on the activity in the venue of interest which will act as a trigger and will be supported, by enhanced video analytics algorithms and techniques which will contribute to decision making procedures”.
The figure below is an update of graphical representation of the P-REACT project concept. This figure was done during the initial phase of the project, when defining the detailed architecture of the P-REACT solution. It includes all the different components/modules that have finally been developed during the project.
In essence, the P-REACT solution is an innovative surveillance system (not yet another “smart” CCTV system- although it is smart). This solution is interoperable (no vendor lock-in), modular, flexible, scalable, dynamically configurable, and last but not least, it is a low cost solution. The P-REACT solution combines multiple-sensor technologies (video, depth, audio) and if focused on prevention and early detection of petty crimes.

Project Results:
The forecasted results and S&T objectives of the P-REACT project are shown in order of relevance and priority:
✓ Objective 1: To design and develop a cost-effective solution for the protection of small businesses and transit locations in an urban environment from petty crimes.
✓ Objective 2: Pioneer research on situational awareness and incidents classification based on the utilisation of privacy-conserving data from the local embedded sensors.
✓ Objective 3: Delivery of novel visual analytics framework for addressing early diagnosis of incidents and enriched offline multimodal visualisation mechanisms towards strengthening decision-making mechanisms.
✓ Objective 4: To design and develop a novel cloud-based service video surveillance as a service (VSaaS).
✓ Objective 5: Delivery of novel low-cost embedded framework with local intelligence and real-time alert categorisation.
✓ Objective 6: To exploit the use of semantic-technologies with evidences from video archive sources in order to forecast, prevent and detect future petty crimes.
✓ Objective 7: To analyse in details technical barriers in the standardisation of the technologies and to ensure that any societal, ethical and legal issues are properly balanced and addressed.
✓ Objective 8: Demonstration, Validation and Evaluation of the proposed system in terms of societal, technological and other key performance and economic indicators.

Potential Impact:
1.1 Potential impact and dissemination and exploitation results
1.1.1 Dissemination plan
The P-REACT Dissemination Plan was developed early in the project lifetime and comprehensively reported in Deliverable D6.2 delivered in M10. Project dissemination has followed four strategic phases as outlined in the figure below.

For the P-REACT project we defined the following groups of interest:
Primary group of interest: This group encompasses the private sector where the impact of petty crime is very high along with law enforcement agencies (LEAs) with the mission to protect society and prevent petty crime incidents. Another group that belongs to this category is private security companies with the mission to provide security services to both the private and public sectors. This group comprises: transportation sector, shop owners, private security companies, law enforcement agencies, and telecommunication companies.
Secondary group of interest: Groups identified as stakeholders in the public sector that may be interested in P-REACT solutions and they can spread the use of P-REACT, such as ministries, city councils, citizen associations and other relevant ones.
Tertiary group of interest: This group includes organisations that their expertise is supplementary to the previous ones, and therefore can influence their operation. Potential groups are researchers, academia, security experts, and organisations currently engaged in related EU or other projects.

1.1.2 Dissemination progress at a glance
With regard to scientific publications, at the time of writing this report two scientific papers have been accepted in conferences and published as part of the proceedings, one more has been submitted and is being reviewed, and one paper is going to be submitted to a journal. In addition, 42 other type of dissemination activities have been recorded, including project presentations, press releases, articles in magazines, interviews in radio stations, or life demonstrations/trials.
The project website was up and running since month 1 of the project and it has been updated regularly. In total the project website has received more than 8,000 visits (sessions) since its creation. In addition, three newsletters plus one special edition have been produced and distributed.
All the aforementioned activities are aligned with the priorities set for the different targeted groups of interest. Section 2.1 of this document includes the list of all scientific and other type of dissemination activities during the project.

1.1.3 End-User Advisory Board
At the beginning of the project three end-user workshops were organised by the consortium members in order to gather from different type of end-user organisations and stakeholders their input which was essential to make a final decision about the use cases and concrete scenarios that were going to be developed during the project.
In the first four months of the project four end-user workshops were organised, one in Athens, another one in Nicosia, another one in Dublin, and another one in San Sebastian. The goal of these workshops was twofold. On one hand these workshops served to gather end-user knowledge and inputs for the preparation of the detailed requirements and use case (scenarios) definition. And on the other hand the workshops were also a very valuable way of raising awareness and interest among stakeholders. Further details can be found in section 4 of deliverable D2.1.
At the time of preparing this final project report, the list of members of the EUAB is as follows: Cyta, Hellenic Police/Forensic Science Division, Trainose S.A PYRSOS Security S.A Railway Procurement Agency (RPA), TPER spa Bolognia, ICTS Hellas, Guardia Municipal San Sebastián, ISS Security S.A. Cyprus Police, Municipality of Bologna

List of Websites: