NA-MIC Project Weeks
Back to Projects List
Make ClassAnnotation Extension Multi-Label and collaborative
Key Investigators
- Lorena Romeo (Magna Graecia University of Catanzaro, Italy)
- Ciro Benito Raggio (Karlsruhe Institute of Technology, Germany)
- Maria Francesca Spadea (Karlsruhe Institute of Technology, Germany)
- Paolo Zaffino (Magna Graecia University of Catanzaro, Italy)
Project Description
ClassAnnotation is an extension designed to support users during medical image annotation process and it is able to generate a structured output for AI applications. Currently, the system does not support multiple label per patient; therefore, it’s desiderable to implement a Multi-Label module to enable this feature. We would also to investigate an approach to make the annotation task collaborative.
Objective
- Create Multi-Label module
- Try to implement a prototype for collaborative annotation
Approach and Plan
- Gather feedback about ClassAnnotation
- Create Multi-Label module
- Discuss about the implementation of a collaboration between annotators
Progress and Next Steps
- Extended ClassAnnotation from single-label to multi-label allowing multiple features per patient (still working in progress).
- Added a clear multi-label user interface with automatically generated class buttons and consistent Single/Multi mode behavior.
- Defined a structured multi-label output.
- Talked to Andras and got some good advice from him.
Next Steps
- Perform thorough testing and validation of the multi-label mode.
- Add support for a collaborative configuration.
Illustrations
