Interests

I am interested in machine learning in general but my research focuses on developing or adapting methods for contexts where data is scarce. Currently, I am interested in deep transfer learning as a way of overcoming data scarcity and I also plan to study interactive learning in the future.

To evaluate and validate the methods, I work with digital pathology problems. Indeed, those are particularly suited for the task as annotated data is often tedious and expensive to obtain in this domain. I am a part of the Cytomine Research team and I work on a day-to-day basis with the Cytomine application and ecosystem (API, data,...).

Publications
ORBi references:
Software and projects
BIAFLOWS

BIAFLOWS is a web based framework to encapsulate reproducibly deploy, and benchmark automated bioimage analysis workflows. BIAFLOWS helps diffusing and fairly comparing image analysis methods, hence safeguarding research based on their results by enforcing highest quality standards.

This is a joint work of many people involved in Work Group 5 of the NEUBIAS COST action including Sebastien Tosi and Raphaël Marée who are leading the project.

Don't hesitate to contact us if you want to contribute to BIAFLOWS with data or workflows, or anything else.

BIAFLOWS application

Preprint on BioRxiv

SLDC

https://github.com/waliens/sldc

SLDC is an open-source Python framework created for accelerating development of large image analysis workflows. It is especially well suited for solving more or less complex problems of object detection and classification in multi-gigapixel images.

The framework encapsulates problem-independent logic such as parallelism, memory constraints (due to large image handling) while providing a concise way of declaring problem-dependent components of the implementer's workflows.

It was initially developed in the context of my master thesis but has been improved since.