France Energies MarinesL'Institut pour la Transition Energétique dédié aux EMR
The postdoctoral candidate will:
- Use and test algorithms that enable the automatic recognition of species that compose biofouling to measure 2 important variables (coverage and roughness) of an image data basis built from work in the ABIOP+ project;
- Assimilate and compare two algorithms developed by UN in order to analyze the advantages of the different algorithms focusing on deep‐learning as an alternative of the initial model based on texture analysis;
- Measure the detection capacity of the algorithms by testing them within virtual settings following the protocol developed by Nantes University and collaborators.
The final objective is to enhance the measure of crucial biofouling variables based on image analysis, especially the roughness of the bio‐colonization, within different environments (e.g. variable lightness, turbidity).
En savoir + : www.france-energies-marines.org/Qui-sommes-nous/Nous-rejoindre