Please note: The algorithm descriptions in English have been automatically translated. Errors may have been introduced in this process. For the original descriptions, go to the Dutch version of the Algorithm Register.
Pressure measurement - group dynamics estimation
Algorithm that classifies movements in a group on camera images
- Organisation
- Municipality of The Hague
- Department
- Stadsbeheer
- Domain
- Field not filled in.
- Algorithm type
- Self-learning
- Status
- Out of use
General information
Name
Pressure measurement - group dynamics estimation
Organisation
Municipality of The Hague
Department
Stadsbeheer
Short description
Algorithm that classifies movements in a group on camera images
Algorithm type
Self-learning
Domain
Field not filled in.
Link to publication website
De camera's zijn toegevoegd aan onderstaande kaart.
https://ddh.maps.arcgis.com/apps/webappviewer/index.html?id=8531785e9a8c4450be8839385003f1bc
Status
Out of use
Application
Purpose
This is a pilot project. The aim is efficient enforcement on crowds. The algorithm is not intended to recognise people or do anything with personal data.
Impact
Citizens are filmed while on the promenade near the information point, after a few seconds the footage is deleted.
Proportionality
It is an initial experiment to determine automated group dynamics rather than physically counting people.
Decision making process
Pressure measurement
Legal
Authorized body
Municipality of The Hague. The police is the data controller, based on 151c. So it is a collaboration.
Formal grounds
Municipal Act. Section 172 Maintenance of public order (camera images) and Section 151C (processing camera images under police direction).
Human Rights Impact Assessment (IAMA)
No
IAMA Description
Field not filled in.
Data Protection Impact Assessment (DPIA)
Yes
DPIA Description
All images are destroyed, the count analysed according to information management standards.
Objection procedure
N/A
Metadata
Contact email
datashop@denhaag.nl
Geographic area
The Hague
Language
Dutch
Use-case
Long description
The algorithm detects people and their speed of movement. The algorithm recognises whether a (moving) object is a human or not and determines the speed for each person. It then assigns a classification to this situation, such as "stationary", "agitated" or "panic".
Usage of base registries
No
Data sources
Camera footage.
Methods and models
Not known to the municipality.
Supervision
Monitoring
The risks of the cameras are tracked. The algorithm itself has no risks.
Human intervention
Yes, the enforcer. Camera and algorithm can be stopped at any time.
Risks
A DPIA has been carried out (DPIA pressure measurement and group dynamics) describing the risks and measures. There is no bias as the dynamics are determined without preference on types of people. The choice of where to hang the cameras was made together with the enforcers.
Evaluation metrics
Signals should be valuable. Evaluations will be conducted. After the experiment, added value will be weighed against importance.