Home   >   Projects   >   FITOPTIVIS

FITOPTIVIS

From the cloud to the edge - smart IntegraTion and OPtimization Technologies for highly efficient Image and Video processing Systems

Field
European
Date
01/06/2018 - 31/05/2021
Industry
  • Agri-food
  • Manufactura
Budget
Total project budget: 22.5 M€ ITI budget in the project: 256,176.81 €.
Funded by

European Commission - H2020 ECSEL

Website
Video

PROJECT INFORMATION

DESCRIPTION

It is a European research project that aims to develop a comprehensive approach for the intelligent integration of image and video processing pipelines for cyber-physical systems, also known by its acronym CPS (Cyber-Physical Systems). Images play a central role in human perception and understanding of our environment. In the same way, SCPs need the visual context to make the right decisions (autonomy) and take the appropriate actions (acting and adapting). However, advanced image and video processing requires handling a large amount of data, while, for adequate performance, results must be available quickly. To meet this requirement, most CPSs feature multiple image processing devices located near (in the front line) of the sensors and actuators. These devices need to operate with low power, as power sources may be scarce or heat dissipation needs to be limited to protect the system, its environment or the humans interacting with it. Balancing power consumption and performance in distributed configurations with a large number of sensors and actuators is a challenge. To address this, FitOptiVis seeks to provide end-to-end multi-objective optimization for image processing pipelines with a special emphasis on power consumption and performance.

It is a European research project that aims to develop a comprehensive approach for the intelligent integration of image and video processing pipelines for cyber-physical systems, also known by its acronym CPS (Cyber-Physical Systems). Images play a central role in human perception and understanding of our environment. In the same way, SCPs need the visual context to make the right decisions (autonomy) and take the appropriate actions (acting and adapting). However, advanced image and video processing requires handling a large amount of data, while, for adequate performance, results must be available quickly. To meet this requirement, most CPSs feature multiple image processing devices located near (in the front line) of the sensors and actuators. These devices need to operate with low power, as power sources may be scarce or heat dissipation needs to be limited to protect the system, its environment or the humans interacting with it. Balancing power consumption and performance in distributed configurations with a large number of sensors and actuators is a challenge. To address this, FitOptiVis seeks to provide end-to-end multi-objective optimization for image processing pipelines with a special emphasis on power consumption and performance.

Impact

Cyber-physical systems have increasingly become self-contained integrations of electronic systems, subcomponents and software, which interact closely with physical systems and their environment.
Some are intelligent systems in themselves. Therefore, new design approaches and methodologies are demanded to deal with this new class of computer systems.
Develop a reference architecture that is capable of describing, from the cloud to the front line, the various networks of heterogeneous devices that together perform image processing in a CPS.
Research devices for image processing that are intelligent, low-power, high-performance, and capable of being integrated into the reference architecture.
Investigate methods and tools to ensure that the different devices that make up a CPS are able to optimize their performance on the fly based on multiple objectives (power consumption, amount of data to be processed, quality of results, etc).

Entities

30 entidades de 5 países

Technological capabilities

Big Data
Data analysis and exploitation