On the 4th of November, we will have Associate Professor Hongli Zeng from Nanjing University of Posts and Telecommunications, China to give us a talk about image segmentation. 

Register to receive the Zoom link:
https://stockholmuniversity.zoom.us/meeting/register/u5Urf--orz8qE9QCrI9SP7yxWCmXAi1lifyn

Abstract:

his talk will present the studies on more than 40 CT scans of guppy brains. The scans were prepared for two groups of guppies, which were selected artificially with large and small brain size defined as the ratio between the length and the weight of the brain. I am trying to answer two questions, which might be interesting from the biological point of view. The first one is the exact sizes of sub-structures for each sample brain while the second is if there are significant differences between two groups of the brain scans. If there were, where they are located. These questions actually bring a critical appraisal of existing neuroimaging techniques as the number of voxels of one sample (about 10^7) is much larger than that of the samples. For the first question, a region-based morphometry (RBM) method is used. For the second one, the mass-univariate techniques are found to be more powerful compared with multivariate (permutation-test and machine learning based) approaches.

 

Prof. Zeng did her PhD at Aalto University and postdoc at Uppsala University.

Her research interests include Complex Systems: Social Dynamics, Topological Models; Inverse Problems: Network Reconstruction, Brain Morphology Analysis of Guppies.

Selected publications:

Maximum likelihood reconstruction for Ising models with asynchronous updates

HL Zeng, M Alava, E Aurell, J Hertz, Y Roudi

Physical review letters 110 (21), 210601

2013
Network inference using asynchronously updated kinetic Ising model

HL Zeng, E Aurell, M Alava, H Mahmoudi

Physical Review E 83 (4), 041135

2011
Evolution of brain region volumes during artificial selection for relative brain size

A Kotrschal, HL Zeng, W van der Bijl, C Öhman‐Mägi, K Kotrschal, ...

Evolution 71 (12), 2942-2951

2017