Skip to main content

Automatic Data relevancy Discrimination for a PRIVacy-sensitive video surveillance

Final Report Summary - ADDPRIV (Automatic Data relevancy Discrimination for a PRIVacy-sensitive video surveillance)

Executive Summary:
The ADDPRIV project developed a new solution to limit the storage of unnecessary data and it proposes technologies for automatic discrimination of relevant data recorded by a multi-camera network, based on the automatic identification of security-relevant events.
In order to achieve this aim, technologies towards a privacy–sensitive video surveillance for public places, with automatic data relevancy discrimination and intelligent storage has been developed. The discrimination of data relevancy required the development of novel algorithms operating on existing multi-camera networks, since relevant data not only corresponds to video scenes capturing individuals´ suspicious behaviour, but also video scenes corresponding to those same individuals recorded by other cameras in the network, so that reconstruction of the route they followed could be performed.

ADDPRIV´s approach is based in the principle that discrimination cannot be implemented at the level of a single camera but at the multi-camera network level as a whole. The suspicious behavior of one individual generates an alert event in one of the network cameras. This alert triggers the instruction to other cameras in the network for automatic detection of that same individual by browsing scenes stored in certain time windows.

The systems developed in the ADDPRIV project, and the solutions offered, were validated and guided by criteria and metrics determined by social and ethical experts and end users. The ethics scoreboard is one stage in system evaluation and another was concerned with ethical compliance.

Demonstration of the system to end users, in particular those who would operate such a system in the future provided an additional pillar for the evaluation.

In terms of technological development, ADDPRIV achieved the integration of a complex privacy enhancing system that enables the exploitation of the opportunities offered by algorithmic technology to deploy a system that determines in a precise and reliable manner private data from video surveillance which is not relevant from the perspective of security or safety and therefore which does not need to be stored. The technology developed in the ADDPRIV project framework permits the automatic discrimination of relevant data recorded on multi-camera networks in order to respect the privacy of individual data on subjects, and it works in compliance with citizens’ privacy rights.

A set of 16 ethical principles has been established in order to define a new ethical assessment for algorithmic video-surveillance systems. These principles have been developed maintaining a constant connection with the technology developments and the solutions developed in ADDPRIV, as well as with respect to the potential ethical opportunities and impacts that have been identified through the ethical evaluation of the technology developments and solutions. These principles have been defined in order to constitute a starting point for assessing video surveillance technologies capable of overwhelming the simple compliance with legally enforceable principles to specially include areas of ethical concerns. However, this set of principles should be considered as a starting point, due to the fact that it needs a continuous process of updating in relation to changes in technology and organizational processes.

Project Context and Objectives:
Video surveillance is extensively used in many European countries for security purposes, contributing to a more effective prevention, detection, and/or prosecution of crime and terrorism. Although its contribution to security is not under debate, its effect on the citizens’ rights has generated a deep controversy. While European nations are committed to respect values of human dignity, freedom and equality for minorities in their use of Security, observation studies carried out in CCTV control rooms have evidenced that video surveillance operations can be subject of discriminatory practices. These studies have shown how ethical minorities and certain social subgroups are much more likely to be the focus of attention of surveillance operators, for reasons that are not explained from their behaviour.

Technologies that can guarantee a well-balanced trade off between security and human rights (equality, freedom and privacy) in the use of video surveillance in public places are needed and are committed to have a major impact on the European societal acceptance of video surveillance. As the Royal Academy of Engineering stated in a 2007 report, “There is a challenge to engineers to design products and services which can be enjoyed whilst their users' privacy is protected. […] Effort should be put into researching ways of monitoring public spaces that minimize the impact on privacy - for example, pursuing engineering research into developing effective means of automated surveillance which ignore law-abiding activities”.

In order to satisfy this need, scientific research must join expertise in different cross-linked areas at a multinational level making possible step changes in the functionality of surveillance, towards a Privacy-Sensitive Video Surveillance. This has been the objective of ADDPRIV: the development of new knowledge and algorithms to build on existing smart video surveillance system, in order to make them comply with the Human Rights European Convention, thus enhancing the quality of living of European citizens as well as the competitiveness of European research and industry.

In fact, the majority of current video surveillance systems fail to be privacy-sensitive. The scientific community focused efforts in developing solutions towards a Smart video surveillance, implementing automatic detection algorithms. Vehicle plates recognition, faces recognition, mobility detection are available in commercial solutions and several R&D projects on automatic detection of abnormal behaviour has been executed or are currently on progress. However, few are the technical examples (projects or products) aimed at minimizing the impact of video surveillance on privacy. Trials have been executed basing tagging of authorised individuals by means of face recognition technologies, RFID tagging, preselected “tag” (i.e. yellow hard hat) recognition or tracking by GPS, in order to mask them electronically protecting their identity.

However, all these different approaches have been conceived for video surveillance in private places, where authorised and non-authorised individuals can be identified and monitored. Thus, these approaches cannot be applied for video surveillance in public places.

ADDPRIV proposed a different approach, building on existing smart video surveillance systems, to address a major need of European society for surveillance systems that comply with citizens´ right to privacy. The project developed and validated new knowledge and solutions to be implemented on existing multicamera surveillance networks.

ADDPRIV tackled this need by developing technologies towards a privacy–sensitive video surveillance for public places, with automatic data relevancy discrimination and intelligent storage. The discrimination of data relevancy is enabled by novel algorithms operating on existing multicamera networks, since relevant data not only corresponds to video scenes capturing individuals´ suspicious behaviour, but also video scenes corresponding to those same individuals recorded by other cameras in the video surveillance network, that allow for a reconstruction of the route they followed. The information that is not discriminated as relevant corresponds to non-useful recordings of law-abiding citizens. Secure deletion technologies have been also developed for erasing non-relevant video files according to the highest security standards.

Taken into account that privacy infringement is due to the fact that other individuals have control over the use and disclose of recorded information where citizens are identifiable, automatic surveillance processes are a major contributing towards a privacy respectful video surveillance. Automatic processes imply not only an automatic detection of suspicious behaviour (smart video surveillance) but also automatic browsing, identification and retrieval of relevant images on the suspicious individuals, recorded before and after the suspicious event and across the surveillance network. ADDPRIV enabled technologies for an automatic discrimination of all recorded data for a given suspicious individual. The developed novel solutions for automatic discrimination of relevant data not only addressed the issue of privacy but also other human rights such as equality and racial discrimination, by eliminating the influence on the operators´ prejudices.

ADDPRIV worked in order to provide a technical feasible solution capable of dealing with the following objective “the generalisation of surveillance systems collecting massive data raises data protection and integrity issues. However, the scope of surveillance could in most of the situations be focused and targeted to critical parts.” ADDPRIV project idea is based precisely on that principle: Video surveillance systems operate continuously, 24 hours, 7days per week, which raises the need for managing enormous volumes of sensitive data. Out of all the information stored, only a very small percentage (less than 1%) is estimated to be somehow useful in crime prevention, detection and prosecution. At this aim, the project developed algorithms for automatic Data Relevancy Discrimination where relevant data not only corresponds to video scenes capturing individuals´ suspicious behaviour, but also video scenes corresponding to those same individuals recorded by other cameras in the video surveillance network, that allow for a reconstruction of the route they followed (where they were before and went afterwards). Those video scenes useful for suspicious individual´s Route Reconstruction are apparently irrelevant, in the sense that they do not show abnormal behaviour; however they are extremely useful for crime prosecution and need to be preserved.

ADDPRIV´s approach is based in the principle that discrimination cannot be implemented at the level of a single camera, on the contrary needs to analyze the multi-camera network as a whole. The suspicious behaviour of one individual generates an alert event in one of the network cameras. This alert triggers the instruction to other cameras in the network for automatic detection of that same individual (features recognition), browsing scenes stored in certain time windows. The development of these solutions has been based on implementing artificial intelligence for learning algorithms, so that they can be rapidly and effectively adapted to different infrastructures and changing situations.

ADDPRIV´s aims to develop solutions for Data Relevancy allowed for Secure Deletion technologies to be applied on the data that was not discriminated as relevant, thus limiting the storage of unnecessary data, corresponding to law abiding citizens. The implementation and validation of these solutions in real life scenarios in Milan Linate Airport and Madrid Viilaverde Alto Commuter station provided results for refining the Data Relevancy Discrimination algorithms in order to implement iterative improvements towards an optimal balance between data storage and data discriminated as irrelevant, susceptible of secure deletion.

Project Results:
ADDPRIV started by precisely defining all legal and ethical specifications that the solutions had to deal with and a preliminarily defining the criteria for the evaluating of system´s compliance with citizens’ privacy rights. At this aim, deep analysis of the requirements of smart video-surveillance systems for better compliance with citizens’ privacy rights has been undertaken: the legal criteria which members of the EU must abide by in terms of data and privacy protection have been precisely defined setting the focus on the countries involved in the ADDPRIV Project: Spain, Italy, UK and Poland. Moreover, the information has been further refined for the consortium end users countries: Italy and Spain. Moreover, the system has been analysed from a social and ethical perspective, setting preliminary criteria for evaluation of system compliance with privacy and ethics requirements. At this aim an Ethics Scoreboard have been developed, setting the basis for the development of new ethical standards for surveillance systems that go beyond legal compliance. Moreover an Ethical Plan has been also developed, presenting in more details all the activities that have been carried out in order to define in detail all the ethical and legal aspects that the ADDPRIV project has to engage. In order to strengthening these aspects, two external boards for the project have been created and maintained: an End Users AB and Ethics AB. The End Users Advisory Board counted with the participation of: Luca Golzi (Italian Polizia di Frontiera), Francisco Prada San Román (Spanish Guardia Civil), Paloma LLaneza (AEDEL – Spanish Association of Electronic Evidence), Francisco Lazaro (Renfe) and Marcella Scuccimarra (SEA Aeroporti di Milano). The Ethics Advisory Board counted with the participation of: Gary Davis (IE DPA), Mª José Blanco and Julián Prieto (Spanish DPA), Leon Hempel (Berlin Technical University), Jan Philip Albrecht (LIBE), Charles Farrier (no CCTV) and Steve Wright (Leeds Metropolitan University).

Once the operational legal and ethics frameworks have been established, the project moved to the technical definition of the solution to be developed. At this aim, a review of existing smart video surveillance systems capable of being integrated with the ADDPRIV project has been performed. It included an overview of the systems available on the market, projects in the field of video surveillance and popular algorithmic methods utilized for particular system components. The work carried out intended to raise awareness of the consortium to the currently utilized approaches for smart monitoring systems and moreover, to identify possible solution which could be applied to the developed system. Finally, technologies developed by GDANSK University in the framework of the INDECT project have been partially integrated. In parallel a detailed description of the end users requirements and a precise scenarios definition have been performed: a detailed definition of the functional requirements of the two end users (SEA and Renfe) and a precise definition of the Pilot Scenarios (Linate Departure Area for SEA and Villaverde Alto Commuter Station for Renfe) have been produced, taking into account flexibility and scalability elements. This led also to the establishment of common criteria on the use of standards to ensure interoperability among the different algorithms and fully define the overall ADDPRIV system design.

Once the technical basis has been established, the technologies for enabling the data relevancy discrimination have been developed. Algorithms for the automatic event detection for triggering the Data Relevancy Discrimination Algorithms have been adapted and integrated in the system architecture. Event detection algorithms integrated in the ADDPRIV project are part of a larger framework developed by GDANSK during the FP7 Project INDECT (GA 218086). Nevertheless, big efforts have been spent in order to adapt the algorithms to the ADDPRIV project requirements. Moreover, according to the end users needs, three algorithms have been developed for the automatic event detection: (a) Luggage abandonment, (b) Counterflow and (c) Intrusion detection. In parallel, knowledge and technologies for Route Reconstruction necessary for reconstruct the path of particular suspicious (triggered by event detection) moving through the monitored area have been developed. For implementing the Data Relevancy Discrimination, Route Reconstruction algorithms have been developed in order to generate a sequence of images that recollects all the visual information related to the event. At single camera level, algorithms for determining the “Single Camera Activity Network (SCAN)”, the parameters defining the individuals’ movements within single camera field of views (FOV) have been developed. Once SCAN technologies have been developed, algorithms for calculating the “Transition Times” (TT) have been developed; these algorithms allow calculating how much time it takes for an individual to exit one FOV and appear in another, constituting, during the multi camera tracking stage, a filtering to generate a considerably smaller candidate list with tracks that only satisfy this property. Basing on SCAN and TT, tracking algorithms that generate a multi hypothesis output with each hypothesis corresponding to an individual’s movement within a camera have been developed. The algorithm executes iteratively to estimate the most likely tracks that belong to the individual in other cameras. As such the output of the algorithm is multi hypothesis and contains a collection of ranked tracks for each hypothesis. An elemental module of the Route Reconstruction system is the object re-identification sub-module. It allows recognizing the objects represented by images from various system cameras, constituting en elemental component for the route discovery. As first part of the Object re-identification technologies, image features extraction techniques have been developed, allowing the parameterization of object appearance in order to acquire the object unified description regardless of its shape, size and etc. This information is then used as basis for the feature matching method developed that allows the recognition of the most similar objects (in terms of appearance).

In parallel with the technologies for Data Relevancy Discrimination, Intelligent Storage technologies suitable for ADDPRIV have been developed. Several strategies for storage have been reviewed, finally settling on a frame-by-frame storage strategy allowing fine-grained relevancy and privacy management, as well as efficient stripping of the data across several storage media to improve efficiency. Access rights management is role-based and tightly related to the privacy status of the content. Also secure deletion technologies to be applied on video surveillance data files according to the highest security standard have been developed. Secure deletion technologies have been developed within the software module “Secure Erase Engine” (SEE). The SEE is composed by three different functional blocks: (a) The “Secure Deletion Scheduler” (SES) in charge of organizing the erasure requests; (b) The “Secure Erase Agent” (SEEA) in charge of the erasure process and (c) The “Log Generator” (SELG) in charge of generating log files certifying that the erasing process has been executed correctly. In order to overwhelm the difficulties of the Secure Erase for SSD (hardware dependent) a three drive strategy has been also developed; it works as follows: (a) 2 different devices alternately used for storing the real time video fluxes received from the cameras network (use time: 24 hours); (b) When the storage device is not used: the video-frames tagged as ‘relevant’ are copied to the third device (permanent storage); the complete device is securely erased by implementing the secure erase procedure included in the ATA COMMAND SET (ASC-2).

Once the technologies have been developed, a final definition of the evaluation metrics and scoreboard for the validation of the ADDPRIV system compliance with privacy has been achieved. The ethics scoreboard was completed following discussions with the ethics board at the first and second ethics advisory board. The scoreboard established 10 legal principles and 6 further ethical principles that the ADDPRIV technology must adhere to. The scoreboard was then used as part of WP6 and the ethical assessment of ADDPRIV and the preliminary results were presented to the final ethics board. In parallel all the ADDPRIV modules have been integrated in the lab and the new UI module has been implemented. Initial integration testing has been performed using video recorded in the airport and starting checking the compliance with the ethics scoreboard. Further integration and testing has been performed in the real environment of Linate airport, collecting video from nine cameras in the arrival area. Several issues have been identified and solved. Continuous testing and improvement of the system has been possible thanks to the remote access granted by SEA to all technical partners. The same activities have been performed in Villaverde Alto Station.

Having the testing results available, the potential impact of ADDPRIV proposed solutions on organisational and operational processes for End Users have been analysed. Interviews have been conducted with personnel from both end user organisations before live testing of ADDPRIV system, to establish current working practices. Subsequent demonstrations of system and collection of feedback from end users on requirements for implementation of ADDPRIV solution have been also performed. In parallel with the functional testing, the potential impact of ADDPRIV proposed solutions on human rights (privacy, equality of minorities) have been also analysed. Four dialogue workshops and twenty-five interviews with representatives from industry, police, potential end users, civil liberty groups, academics and regulators, along with an extensive review of relevant documents have been conducted, in order to analyse the current ADDPRIV technology effect and identify future development paths.

Potential Impact:
In the framework of the ADDPRIV project, the potential impact can be analysed both in terms of effect on individual privacy and ethics and in terms of operational processes in end users organizations.

As per the first, a comprehensive assessment of the potential ethical opportunities and impacts identified through the ethical evaluation of the ADDPRIV technology and solutions have been performed. In order to explore the possibility of developing ethical practice in the use of video surveillance, and in particular ‘smart’ video surveillance, which can extend beyond what is written in law, the research team worked with a series of research participants, from a range of experts in the fields of privacy, surveillance, security, law, regulation and ethics, as well as members of the public. This approach, exploratory in nature, established a set of possible ethical opportunities that the ADDPRIV project has created. The project also documented the ongoing challenges that persist in the development of this particular privacy sensitive video surveillance system. The potential ethical impacts of the proposed ADDPRIV technology and solutions that have been identified are: Privacy; Surveillance; Security and Safety; Accountability; Mediation; Efficacy; Communication; Risk; Resistance; and Sorting. For any of them, the ADDPRIV technology and proposed solutions do work toward identifying and addressing potential ethical impacts of a ‘smart’ or algorithmic video surveillance system. The project has worked to identify and so distinguish between different ethical problems that are put forward as important considerations when considering the ethics of surveillance and in the context of ADDPRIV’s work to develop a more ethical ‘smart’ or algorithmic surveillance system. The project activities has shown that reducing ethics to privacy is overly simplistic and can not account for the different ethical problems and opportunities that emerge in and through an ethical examination of a surveillance system. As well as proposing a set of potential ethical impacts, the project activities demonstrate the importance of moving beyond privacy in discussions around the ethics of surveillance. It has been also highlighted the need to move beyond legal discussions of surveillance ethics, it promotes the ongoing investigation of the ethical dimensions of surveillance, and examining the potential and development of ethical standards for ‘smart’ video surveillance systems. The methodology designed through the course of this research has enabled an experimental and exploratory engagement with the ethical impacts of ADDPRIV, contributing to the development of the technology solutions. The methodology has gone some way to offering a framework that can ensure an ongoing ethical evaluation. This is significant for two reasons, firstly as the project is situated within the shift toward studying ethics as an on-going set of practices, and secondly as ADDPRIV is committed to developing an ethical assessment for a ‘smart’ or algorithmic video surveillance system. This novel approach is different to more traditional consequentialist or deontological theories of ethics that are commonly the basis for assessments of the ethics of surveillance. This is important as there has yet to be a developed research programme working through the likely ethical consequences or appropriate normative principles for video surveillance. Ethical issues may vary depending on circumstance and on the ways a system is set up or used, a single set of principles is not deemed appropriate here. An ongoing approach to studying and analysing the ethical implications of algorithmic video surveillance is more appropriate. Although the ten potential ethical impacts reported reflect the research conducted for ADDPRIV, the report can also be used as a stand-alone document for others to use who have an interest in developing and testing ethically-oriented ‘smart’ or algorithmic surveillance systems.

In terms of operational processes for end users organizations, ADDPRIV developed a system for use in CCTV networks which would reduce the impact of those networks on the privacy of individuals whose images are captured in surveillance footage. The project solution achieved this by detecting events which are relevant to security and mark all related footage. ADDPRIV makes use of a route reconstruction component to assemble footage which contains the subject or object of interest. Once the operator confirms whether or not marked footage is relevant to security, the footage that is irrelevant to security can be deleted, thereby reducing the amount of footage that is retained. The ADDPRIV project has developed a new solution to limit the storage of unnecessary data and it proposes a solution for automatic discrimination of relevant data recorded by a multi-camera network, based on the automatic identification of security-relevant events. The events considered are: (a) Left luggage; (b) Barrier crossing (intrusion in a forbidden area) and (c) Counter flow. Abandoned luggage represents a potentially dangerous situation for public safety, especially in public places, such as at airports. Barrier crossing detection covers two situations 1) when an object/person enters a restricted area or walks into an unauthorized zone and 2) where people can be present on both sides of an area, but crossing the area is forbidden. Counter flow is triggered when an object/person moves in an opposite direction to the general flow. The systems developed in the ADDPRIV project, and the solutions offered, were validated and guided by criteria and metrics determined by social and ethical experts and end users. The ethics scoreboard is one stage in system evaluation and another was concerned with ethical compliance. Demonstration of the system to end users, in particular those who would operate such a system in the future, provided the third pillar of evaluation.

The ADDPRIV solution offers an innovative approach through the implementation of technologies for a privacy-sensitive video surveillance in public places, with automatic data relevancy discrimination and intelligent storage according to images relevancy. Images considered by the system as not relevant (not associated to security sensitive events) are deleted according to the highest security standards, in order to ensure that privacy rights are preserved during the lifetime of the data and after data erasure, minimizing the impact on citizens’ privacy. Moreover, by strongly limiting the intervention of CCTV operators, consequences such as having ethnical subgroups more likely to be the focus of attention of surveillance operations are eliminated, contributing to the respect of values of human dignity, freedom and equality for minorities.

The technology developed in the ADDPRIV project are compatible with existing multicamera video surveillance networks and can be relatively easily integrated within them. Once deployed, the system will be able of automatically detects potentially suspicious situations and of generating a sequence of images that recollects all the visual information related to the event. Thanks to the introduction of the concept of intelligence applied to the complete cameras network, ADDPRIV allows the correlation of information for different sensors, allowing an automatic, accurate and reliable determination of what is and is not relevant information from a security perspective. Any irrelevant information is subsequently deleted.

ADDPRIV adopts a highly synergic approach with ethics and privacy: it constitutes an effective video-surveillance system designed to respect citizens’ right to privacy through the use of: (a) Efficiency indicators – determined by organizations responsible for the management of video-surveillance systems; (b) Social and ethical impact indicators – determined by data protection and privacy experts, and ethical experts working in the area of human rights and civil liberties, in the form of an ethical framework defining specific criteria that the system must adhere to in terms of data and images storage and selection.

The ADDPRIV solution can be used for enhancing the security and public places, due to the improved efficiency and reduced reaction time in critical situations for organizations: (a) Automatic detection of security sensitive events (b) Automatic collection of the visual information proceeding from all the sensors in the network, related to the security event detected. It also enables the reduction of the storage capacity required for the videosurveillance system operation: compliance with the principles of personal data recollection minimization.

ADDPRIV has a strong potential for improving the social acceptability of videosurveillance technologies, due to the inclusion of the value and importance of privacy alongside enhanced security.

Organisations and system operators can gain in efficiency and achieve increasing the customers satisfaction by meeting the expectation of maintaining security required level without an intrusive instrument for their private life sphere.

Video surveillance is extensively used in many European countries for security purposes, contributing to effective prevention, detection, and/or prosecution of crime and terrorism. Although its contribution to security is not under debate, its effect on the rights or citizens has been the subject of much controversy. While most countries are committed to respecting values of human dignity, freedom and equality for minorities, studies in CCTV control rooms have shown that video surveillance operations can be the subject of discriminatory practices.

The backdrop for the ADDPRIV project was the need for new technologies for video surveillance in public places that can guarantee a well-balanced trade-off between security and privacy protection. Existing video surveillance systems fail to be privacy-sensitive. The focus of the scientific community has been on developing smart video surveillance with much fewer efforts or focus on finding ways to minimise the impact of video surveillance on privacy.

The ADDPRIV project tackled this need by developing technologies towards a privacy–sensitive video surveillance for public places, with automatic data relevancy discrimination and intelligent storage. The discrimination of data relevancy required the development of novel algorithms operating on existing multi-camera networks, since relevant data not only corresponds to video scenes capturing individuals´ suspicious behaviour, but also video scenes corresponding to those same individuals recorded by other cameras in the network, so that reconstruction of the route they followed could be performed.

ADDPRIV´s approach is based in the principle that discrimination cannot be implemented at the level of a single camera but at the multi-camera network level as a whole. The suspicious behaviour of one individual generates an alert event in one of the network cameras. This alert triggers the instruction to other cameras in the network for automatic detection of that same individual (features recognition) by browsing scenes stored in certain time windows. The development of these solutions are based on implementing artificial intelligence for learning algorithms, so that they can be rapidly and effectively adapted to different infrastructures and changing situations.

ADDPRIV value proposition for video-surveillance systems manager:
- ADDPRIV allows for more efficiency and reduced reaction times in security sensitive situations for organizations thanks to the combination of technologies for automatic event detection and for the automatic collection of the visual information related to,
- ADDPRIV brings together the benefits of really enhanced security with a Privacy by design videosurveillance technology, improving existing technologies with a new focus: shifting from intelligence applied on individual cameras to an integrated use in a distributed sensors network
- ADDPRIV provides a way to reduce the storage capacity required for the video-surveillance system operation, easing the compliance with the principles of personal data recollection minimization
- ADDPRIV provides organizations with a technology with potential to improve the social acceptability of videosurveillance technologies, due to the inclusion of the value and importance of privacy alongside enhanced security

ADDPRIV value proposition for citizens:
- ADDPRIV is an objective system where decisions are placed on technology, rather than human, limiting the potential for targeted selection and discrimination based on physical characteristics and traits
- The Route Reconstruction features provided by ADDPRIV enable a significant reduction in the amount of individual data collected and viewed by security operators, limiting citizens’ feeling of invasion into their private sphere in order to maintain the security objectives established by the organization.
- The combination among the Event Detection, Route Reconstruction and Privacy Enhancement modules have a great potential for greater social acceptance of video-surveillance: implementation of the principle of data minimization (storage of the minimum amount of personal data/images required for achieving the security objectives) will contribute to reducing privacy concerns. This will ensure secure and timely deletion of irrelevant data and images.

The project consortium worked in order to raise awareness around the project activities troughout all the project execution. The consortium has conducted extensive dissemination on the ADDPRIV project and has produced substantial levels of output across different media types. Publications have been made in technical and socio-technical conference and journals, have made presentations in workshops and meeting forums, have shared information with other FP7 projects working in similar fields, made contributions to public media and engaged in local debates about privacy and security. In the Description of Work, the consortium proposed to publish a total of 41 publications as either scientific papers, book chapters, conference papers and contributions to technical events and exhibitions. In the first 2 years of the project, the consortium had published 29 of these types of publications and, in the last year, they published a further 16, giving a total over the lifetime of the project of 45. Moreover further journal and conference papers are currently under review or in progress so this total publication output will rise still further.

List of Websites:

www.addpriv.eu