digeiz.

The digeiz team works as three departments :

• the SALES identifying customers and discussing how footfall insights will transform their core business for shopping malls, train stations and airports

• the OPS responsible for defining the camera scenarios (on-map and on-site studies) and then conducting the hardware deployment (cameras, servers, network, ...) and its calibration

• the TECH developping the overlay of AI for videoprotection cameras, as well as the infrastructure to store these data and the products to maintain the entire stack

All teams are heavily relying on each other to foster our product features and increase their business impact.

The SALES sign deals with our customers.

Our solution implies a complete transformation of our customer business : turning a shopping mall into a media - a way to broadcast ads, measure the audience and monetize it. Therefore, they usually have discussions with the C-levels and investors for these companies.

The solution is interesting for them if it is deployed large shopping malls, and in many of them. Shopping malls is the start to improve the technology, we also aim any large crowded area such as train stations, airports, stadiums...

The OPS play a crucial role in the complete lifecycle for deploying a new site :

• Pre-sale : we make on-map quotations and propose relevant camera scenarios for the SALES team

• Study : we go on-site, study maps, make pictures where all the cameras should be installed, and fill all the infos inside our dedicated google-maps like platform for the third-party installers

• Reception : we validate the camera once installed, check connectivity, stream quality, and calibrate for counting

• Maintenance : we check data availability, camera health status and open tickets with the third-party to make sure everything is up and running

The TECH team is split into several departments :

• R&D is responsible for prototyping the algorithms

• The C++ team turns into them into industrial software which generates the footfall data from the camera video streams

• The python backend receives and stores the data into no-sql databases

• The frontend creates the web UI for displaying the mall configuration and the interface for the maintenance tools

We develop algorithms for generating data (counting, gender/age classification, reidentification) as well as maintenance and monitoring tools (dataset drift, retraining, anomaly detection, ...). The final data is relevant only if the whole pipeline is working.