background

Work Plan

Work Packages Overview

WP1: Multi-Stakeholder Mapping and Requirements Analysis
  1. Identify relevant SSA stakeholders throughout the project.
  2. Conduct qualitative studies to identify ELSI, infrastructure, healthcare, policy, and training requirements.
  3. Organise co-creation workshops to refine the multi-stakeholder requirements.
  1. Establish the AFRICAI-RI platform for FAIR African medical imaging data storage.
  2. Develop a one-stop-shop data anonymisation, curation, and upload toolkit.
  3. Create a catalogue facilitating systematic searching of African imaging datasets.
  4. Establish the AFRICAI-RI federated learning platform for accessible medical imaging AI in Africa.
  5. Create a multilingual integrated web portal integrating all components.
  1. Establish operations for FAIR and secure African data management.
  2. Establish a scalable legal infrastructure and governance for African (federated) medical image AI.
  3. Collect and curate data of 10 SSA countries in the AFRICAI-RI platform.
  4. Establish an easy-to-use data access portal and user manager.
  5. Facilitate scalability to the varying African settings by providing AFRICAI-RI onboarding protocols.
  6. Perform rigorous stress-testing and validation protocols of the AFRICAI-RI platform.
  1. Implement a trustworthy AI toolkit for diagnostic imaging across SSA.
  2. Define criteria and best practices for trustworthy AI validation in SSA contexts.
  3. Implement AI tools for each of the project’s use cases by leveraging the multi-centre data from WP3.
  4. Gather feedback from the use cases to improve all tools from WP2 and WP3.
  1. Organise open calls to assess the AFRICAI-RI infrastructure’s security, usability and scalability.
  2. Invite clinical sites across SSA to join the AFRICAI-RI infrastructure as new data nodes.
  3. Invite AI research teams and companies across SSA to develop new AI solutions based on new datasets.
  4. Gather feedback to enhance the infrastructure’s functionalities and security measures.
  1. Deploy the AI tools from UC2 and UC3 in SSA sites for clinical, feasibility and usability evaluations.
  2. Implement two in-silico clinical trials to assess clinical value and applicability in SSA of (1) AI-powered diagnostics for paediatric CXR and (2) AI-enabled CUS in adults.
  3. Assess the feasibility of a continuous federated learning system to enable decentralised learning, with updates and improvement of the CAD4TB model over time.
  1. Develop multi-disciplinary training modules tailored to diverse stakeholders across SSA.
  2. Organise a set of training events, including monthly 2-hours webinars and two in-person schools.
  3. Establish structured mentorship and early career research programmes across SSA.
  4. Provide all training materials and resources in diverse languages to enhance accessibility.
  1. Perform an a posteriori socio-ethical analysis of the digital technologies for AI-assisted radiology in SSA.
  2. Enhance the credibility of the AFRICAI-RI infrastructure through a data protection impact assessment.
  3. Develop a long-term sustainability plan for AFRICAI-RI’s infrastructure.
  4. Define pathways for the integration and adoption of the federated AI tools into SSA’s healthcare systems
  1. Oversee the effective execution of research activities (WP1-WP8), while ensuring adherence to established timelines, budgets, and quality standards, and actively monitoring risks and mitigation measures.
  2. Continuously disseminate the project’s progress and outcomes to clinical, research, and industrial stakeholders, as well as to authorities and the general public.
  3. Foster interactions and collaborations between related projects in SSA and internationally
  1. Coordinate and manage all aspects related to project administration, finances, and contractual obligations.
  2. Continuously communicate the project’s progress and outcomes to clinical, research, and industrial stakeholders, as well as to authorities and the general public.
  3. Develop a comprehensive plan for IP management and exploitation of the project’s key exploitable results.
Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors