Machine Learning for Biological and Medical Imaging
ApplyProject Description
We have recently developed hybrid machine learning techniques for retinal images. For publications see google scholar(https://scholar.google.com/citations?hl=sv&user=_DUppAgAAAAJ&view_op=list_works&sortby=pubdate). Challenges include limited number of images, unbalanced data-sets, and interpretability of feature representations. Subprojects include to
Formulation and training of robust generative models (e.g.GANsand versions thereof) for the Retinal Dataset-
Extend and apply the techniques to melanoma datasets Develop and apply techniques to identify meaningful (biological/medical) feature representation from a successfulclassification