This project aims to explore data processing techniques, namely missing data, in order to merge information from public datasets. The datasets may contain metadata, such as age, gender, or location of the lesion as well as clinical and dermoscopic images annotated for different skin diseases. The merged information will then be used to build a multimodal model for skin disease classification.
> Expand AICOS background knowledge in: 1) Multimodal Data pre-processing techniques; 2) Multimodal Machine Learning;
> Expected direct impact in current and future AICOS Derma portfolio projects.
Author: David Oliveira
Type: MSc thesis
Partner: FCUP – Faculdade de Ciências da Universidade do Porto