Quividi does not collect or record any image or video, it only collects the anonymous “metadata” that describes the size and the demographics of an audience. As soon as our video processing solution processes an image, the image is immediately ‘forgotten’ by the software.
The Quividi software does not perform face recognition, so it does not have the ability to recognize an individual, either in absolute terms (full identity) or in terms of repeated exposures (e.g. recognizing that someone was at a sequence of different locations, or visited a same location twice). The software can only determine if some anonymous individual is looking at a given interest point, for how long, estimate their basic demographic characteristics such as gender and age, and mood. While Quividi strives to improve the range of attributes that it can use to qualify an audience, it will never produce uniquely identifiable data.
For each detected watcher, the information collected by the Quividi software includes:
- the estimated gender (male or female)
- an estimate of the age (either in bracket or in absolute value)
- the estimated presence of glasses, beard and moustache
- the estimated mood (from very unhappy to very happy)
- the total dwell time while in the field of view of the video sensor
- the total attention time (i.e. the time when a person’s face is turned towards the video sensor)
- the position and distance from the video sensor
In addition to the above data, which is collected for each individual viewer, Quividi produces an estimate of the total footfall in front of the video sensor.
Quividi’s software relies on face detection and face analysis, not on face recognition. In other words, while our algorithms can detect the presence of a person and can determine some broad characteristics such as gender and age, they cannot determine the identity of a person, either in absolute terms (i.e. by associating a face to an identity record) or in relative terms (i.e. by recognizing the recurring appearance of a face at different measuring points). All data is processed in real time and no visual or biometric information is stored in long-term memory or uploaded to the cloud.
Additionally, all the anonymous audience measurement data generated by our software is securely encrypted and securely uploaded to our servers; it is never sold or shared with third-party entities and only the legitimate owners have access to it.
The anonymization takes place locally, on your player, where the Quividi software (VidiReports) is running.
The system is “anonymous” by design since, to start with, VidiReports never records at any point in time any personal data or any image/video of any face. Also no human person ever sees an image while in operation (a setup and maintenance mode exists to help locally verify where the video sensors is aiming at, but it is not used in operation)
Yes. Quividi’s products adhere to the most common privacy laws and regulations wolrdwide. Since we do not collect uniquely-identifiable metrics, our audience measurement figures are not considered “personal data” and therefore no special provisions should be made before installing Quividi’s products.
Yes, Quividi complies with EU current regulations with regards to Data privacy.
The key point is that the data generated by Quividi’s software cannot be considered “personal data”, as per the definition that can be found in various pieces of legislation. For instance, “personal data” is defined by the European Directive 95/46/EC as “information relating to an identified or identifiable natural person (“data subject”); an identifiable person is one who can be identified, directly or indirectly, in particular by reference to an identification number or to one or more factors specific to his physical, physiological, mental, economic, cultural or social identity;” (art. 2 a).
We are aware, though, that the legislation might vary country by country.
If there is no legislation with regards to AVA in a country, we suggest our customer to comply to the recommendations of the Working Paper on Intelligent Video Analytics produced by the International Working Group on Data Protection in Telecommunications.
We are also aware that a new EU regulatory framework – called GDPR – will apply next year (from 25 May 2018), unifying data protection laws across the 28 European Union members.
Quividi has started exchanging with local GDPR experts and we are confident that within this new framework Quividi will be able to legally, fully and more easily operate throughout Europe.
Quividi is the owner of the anonymous audience and traffic data; it licenses it to its customer.
Quividi also processes it on behalf of its customers, using automated methods, in order to extract insights from it.
When a company equips its network with Quividi’s solution for audience measurement, only this company has access by default to the collected data via a secure web-based interface where the encrypted anonymous audience and traffic data is turned into insightful graphs and tables.
This company can ask to have an access prepared on behalf of third parties it trusts, such as an integrator or an agency. Those 3rd parties are then granted separate (generally view-only) access to the online dashboards, or to some reports prepared in VidiCenter.
It can also instruct Quividi to grant API access to its own data on behalf of other third parties, such as SSP (Supply Side Platforms) or DMP (Data Management Platforms), for them to build additional services.
At last, Quividi’s back office support staff is authorized to access the central database for maintenance and troubleshooting reasons; all staff are all held to a strict non-disclosure agreement and a high privacy standard.
The information about an audience’s interest level provided by Quividi allows marketers and advertisers to best respond to consumers’ preferences in a non-obtrusive fashion; consumers are no longer inundated with marketing material that bears little use or interest but “feel” the advertising landscape as more in tune with who they are.