DermImageModality – Image modality classification in dermatology


The goal of this project is to develop a model that can accurately differentiate skin image types. The model should also classify if the image has a ruler or not, since it might be later used to infer the size of the lesion. To accomplish this goal, several machine learning or deep learning approaches will be explored. The developed models should also be able to receive new information through the usage of incremental learning techniques. Transfer learning techniques might also be used to improve the model performance.


Expected direct impact in task execution of current and future Computer Vision projects, such as DermAI. Additionally, it will also contribute to the dissemination indicators of the referred funded project.


Author: Ana Morgado

Type: MSc thesis

Partner: FEUP – Faculdade de Engenharia da Universidade do Porto

Year: 2021