ANC Seminar - 09/06/2020



Arno Onken


Prof Sotirios Tsaftaris


Doing more with less by disentangled data representation


Healthcare is under a perfect storm with AI/ML being offered as a solution to relieve at least one bottleneck: data analysis. Indeed, the detection of disease, segmentation of anatomy and other classical image analysis tasks, have seen incredible improvements due to deep learning. Yet these advances need lots of data: for every new task, new imaging scan, new hospital, more training data are needed. Amongst the vast body of our work, for this talk, I will focus on disentangled representations and how they can help in healthcare and what particular challenges healthcare applications introduce. I will present a framework of disentangled mixed-dimension tensor embeddings (similar to content-style in computer vision) suitable for several analysis tasks that can do more (tasks) with less (supervision). Within a multi-task learning setting this framework can learn embeddings drawing supervision from self-supervised tasks that use reconstruction and also temporal dynamics, and weakly supervised tasks from health records. I will then present a challenging task on multi-modal (multi-view) learning and inference for segmentation based on a disentangled framework. I will use this to discuss how different inductive biases affect disentanglement. Finally, time permitting, I will briefly show recent work on taking disentanglement a step further: disentangling pathology from normal anatomy, and also disentangling the effects of ageing (without relying on longitudinal data). I will conclude by highlighting opportunities and challenges in learning disentangled representations in healthcare and in general.

Papers that will be discussed (in approximate order):

1. A. Chartsias, T. Joyce, G. Papanastasiou, S. Semple, M. Williams, D. Newby, R. Dharmakumar, S.A. Tsaftaris, 'Disentangled Representation Learning in Cardiac Image Analysis,' Medical Image Analysis, Vol 58, Dec 2019

2. G. Valvano, A. Chartsias, A. Leo, S.A. Tsaftaris, 'Temporal Consistency Objectives Regularize the Learning of Disentangled Representations,' First MICCAI Workshop, DART 2019, in Conjunction with MICCAI 2019, Shenzhen, China, October 13 and 17, 2019.

3. A. Chartsias, G. Papanastasiou, C. Wang, S. Semple, D. Newby, R. Dharmakumar, S.A. Tsaftaris, Disentangle, align and fuse for multimodal and zero-shot image segmentation,'

4. T. Xia, A. Chartsias, S.A. Tsaftaris, 'Pseudo-healthy synthesis with pathology disentanglement and adversarial learning,' to appear Medical Image Analysis.

5. T. Xia, A. Chartsias, S.A. Tsaftaris, 'Consistent Brain Ageing Synthesis,' MICCAI 2019. 6. T. Xia, A. Chartsias, C. Wang, S.A. Tsaftaris



Canon Medical/Royal Academy of Engineering Research

Chair in Healthcare AI Chair in Machine Learning and Computer Vision at the University of Edinburgh (UK)

Turing Fellow Alan Turing Institute

web: // email: // twitter: @STsaftaris Photo:


Jun 09 2020 -

ANC Seminar - 09/06/2020

Prof Sotirios Tsaftaris