DisADis

DISADIS project focuses on the challenge of technological solutions for the detection and identification of individuals, proposing the development of a system with distributed deep learning over 5G networks It is a goal of the first phase of the project to build an entire E2E system of person recognition, using fiber communication, leaving 5G communications for the second phase of the project. Although there are solutions in this field on the market, there are significant technological challenges related to the computational load of the solutions (which requires a high-cost infrastructure) and the low effectiveness in the recognition of low-resolution images, among others. To achieve this, and as a first step, all technical and performance requirements are designed.

The DISADIS project focuses on detection and identification of individuals, proposing the development of a system with distributed deep learning on 5G networks. The following use cases are related to face detection and face recognition Machine Learning (ML) functionalities. Although one may thing that those ML problems are already solved, there are still many use cases in which such algorithms are not robust enough due to many uncontrolled conditions and variability or end user devices cannot afford the computational cost associated to them. The specific use cases that will be validated during the execution of the process, which are all focused on access control areas. 

The technology behind these use cases is envisioned as dual technology, since its final application covers both the civil sphere (access control, audience identification at events, traffic counts in companies, buildings, etc.) and in the field of security and defense (identification of subjects in search and capture, identification of people of interest in public places, etc

Funded Project

CDTI Centro para el desarrollo Tecnológico Industrial del Ministerio de Ciencia e Innovación IDI-20210021
 

The DISADIS project is development of NOKIA