UCL Centre for Medical Image Computing


Georgios Lazaridis - Fabio Ferreira - joint CMIC-WEISS seminar series

25 September 2019, 1:00 pm–2:00 pm

Georgios Lazaridis - Fabio Ferreira - a talk as part of the joint CMIC-WEISS seminar series

Event Information

Open to



cmic-seminars-request@cs.ucl.ac.uk – UCL Centre for Medical Image Computing


Function Room, 1st floor
90 High Holborn
90 High Holborn

Georgios Lazaridis - Fabio Ferreira

Fabio Santos Ferreira

Title: Multi-view machine learning approaches for mental health disorders

Abstract: Current scientific research, evaluation, and treatment in psychiatry are based on signs and symptoms rather than objective biomarkers of illness. It has been shown that the diagnostic categories do not align with findings emerging from clinical neuroscience and genetics, and fail to predict treatment response. This opens a window of opportunity for more exploratory machine learning approaches, such as Canonical Correlation Analysis, Partial Least Squares and Group Factor Analysis. These methods can be used to uncover meaningful multivariate relationships between different sources of data (e.g. brain imaging and behaviour assessments) and potentially find reliable subgroups of patients. I will present some applications of these methods and how they can be used as building blocks of more complex hierarchical models which might ultimately help tackling the heterogeneity present in mental health disorders.


Georgios Lazaridis

Title: Enhancing OCT Signal by Fusion of GANs: Improving Statistical Power of Glaucoma Clinical Trials

Abstract: Accurately monitoring the efficacy of disease-modifying drugs in glaucoma therapy is of critical importance. Albeit high resolution spectral-domain optical coherence tomography (SDOCT) is now in widespread clinical use, past landmark glaucoma clinical trials have used time-domain optical coherence tomography (TDOCT), which leads, however, to poor statistical power due to low signal-to-noise characteristics. Here, we propose a probabilistic ensemble model for improving the statistical power of imaging-based clinical trials. TDOCT are converted to synthesized SDOCT images and segmented via Bayesian fusion of an ensemble of generative adversarial networks (GANs). The proposed model integrates super resolution (SR) and multi-atlas segmentation (MAS) in a principled way. Experiments on the UK Glaucoma Treatment Study (UKGTS) show that the model successfully combines the strengths of both techniques (improved image quality of SR and effective label propagation of MAS), and produces a significantly better separation between treatment arms than conventional segmentation of TDOCT.