Tender
Digipolis has marketed phase 1 (anonymisation of imagery) and phase 2 (image recognition and algorithms for insights) of ALTRUIS combined in one engagement, shifting the technological scope from 'Graphics Processing Unit (GPU)-based AI (in the cloud or on premise)' to 'AI at the edge (edge-AI)'. Specifically, instead of bringing sensor data to the cloud and processing it there with AI (which is inadequate for systems that need to be able to decide/intervene in milliseconds), AI will be brought to the sensors in order to make the Internet of Things (IoT) truly 'smart'.
The contract tendered by Digipolis Development of basic infrastructure for image recognition with edge AI included the following:
- Setting up a Proof Of Concept (POC) with the basic infrastructure for the analysis of (camera) images on edge devices;
- Developing two use cases: (1) (pseudo)anonymisation of images, and (2) object detection and classification applied to traffic measurements.
The assignment was not yet focused on hardware development but only on formulating advice on the specifications of the required hardware (including the edge devices) and a proposal for the overall (edge computing) architecture picture.
The contract was tendered through Digipolis' proven Buy from startups procurement procedure. This is a 'light' procurement approach specifically aimed at involving small(er) innovative market players in Digipolis' procurements and combines the Negotiation Procedure Without Prior Notice (Negotiation Procedure without Prior Notice) with some aspects of Pre-Commercial Procurement (PCP). Buy from startups is a two-stage procedure that first selects suitable candidates for a contract and then conducts negotiations with the selected candidates who have submitted bids. Those negotiations result, through a Best and Final Offer (BAFO) or otherwise, in the award of the contract.
The contract for the development of basic infrastructure for image recognition with edge-AI was awarded to the supplier Rombit NV.