Multi-modal data sources and few-shot learning in dermathology

External Member

David Ahmedt-Aristizabal

Description

Abstract: This project considers the problem of clinical image classification for the purpose of aiding doctors in dermatological disease diagnosis. The system must rapidly generalize to diagnose novel conditions given very few labelled examples with huge intra-class variability using different data sources (intra- and inter-subject variability and variations of dataset and annotation). Dataset: public datasets (e.g. SD-198, Derm7pt).

 Contact: David Ahmedt-Aristizaba

Updated:  1 June 2019/Responsible Officer:  Dean, CECS/Page Contact:  CECS Marketing