DCM is designed from the ground up to provide crowd safety and comfort analytics with minimum impact to the privacy of individuals.
This is a significant difference between DCM and other crowd analytics software.
In summary:
-
No use of facial recognition technology to identify individuals.
-
Mood analysis is done by 'pattern matching' 'smiles and frowns' and results aggregated at a crowd level. DCM’s application-as-a-service is designed to measure and report and enable a crowd manager’s response only as to crowd (aggregated) sentiment, not the mood or sentiment of any individual within that crowd.
-
Camera feeds are sampled via a secure VPN link, with data encrypted in transit and at rest.
-
Data does not leave the relevant jurisdiction - in the case of EU for example, its stored in a European AWS instance.
-
Only the sample video frames used on the dashboard for context are stored for any length of time.
-
Those images can be auto deleted at a customisable time frame negotiated with the customer. Data can also be deleted - but since this is completely anonymous and aggregated at a crowd level it is usually not required by most customers.
-
Video collected at customer events is not used to train the models.
-
There is an option to blur displayed images on the dashboard in real time.
-
There is also an option to turn off model analysis for customers particularly concerned about mood analysis - eg to collect density / headcount only - which is sufficient for certain use cases.
-
While cloud deployments are the preferred implementation type, there is an 'on premises' option for customers who wish for camera footage to not leave their local network.