Automated segmentation and motion correction of the fetal brain Kevin Keraudren Imperial College London December 6 th , 2013
Automated segmentation and motion correction of the fetal brain
Jun 27, 2015
These slides are associated with the following NeuroImage publication:
Keraudren, K., Kuklisova-Murgasova, M., Kyriakopoulou, V.,
Malamateniou, C., Rutherford, M.A., Kainz, B., Hajnal, J.V., Rueckert, D., Automated
Fetal Brain Segmentation from 2D MRI Slices for Motion Correction, NeuroImage (2014),
doi: 10.1016/j.neuroimage.2014.07.023
http://www.sciencedirect.com/science/article/pii/S1053811914005953
For more details on using an XKCD style for latex beamer, visit: http://kevin-keraudren.blogspot.co.uk/2014/03/xkcd-style-beamer-presentation-latex.html
Keraudren, K., Kuklisova-Murgasova, M., Kyriakopoulou, V.,
Malamateniou, C., Rutherford, M.A., Kainz, B., Hajnal, J.V., Rueckert, D., Automated
Fetal Brain Segmentation from 2D MRI Slices for Motion Correction, NeuroImage (2014),
doi: 10.1016/j.neuroimage.2014.07.023
http://www.sciencedirect.com/science/article/pii/S1053811914005953
For more details on using an XKCD style for latex beamer, visit: http://kevin-keraudren.blogspot.co.uk/2014/03/xkcd-style-beamer-presentation-latex.html
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Automated segmentation and motioncorrection of the fetal brain
Kevin KeraudrenImperial College London
December 6th, 2013
Problem
As the fetuse moves
the data is acquired as stacks of misaligned 2D slices
2/12
Problemorthogonal
stacks of 2D slices3D volume
3/12
Definitions
recall=Volume of correctly classified voxels
Ground truth volume of the brain
recall=TP
TP + FN
precision=Volume of correctly classified voxels
Detected volume of the brain
precision=TP
TP + FP
4/12
Idea
99%
91%
96%
93%
Box detection
RF/CRF segmentation
Enlarged segmentation
Final segmentation
0
20
40
60
80
100
recallprecision
5/12
Box detection
MSERSize
filteringBag-of-SIFT RANSAC
MSER regions Filtering by size andBag-of-Words
6/12
Box detection
7/12
Brain extractionRandom Forest classifier
and Conditional Random Field
8/12
Motion correction
Acquisition ofsnapshot images
Automatedmasking
Volumeregistration
Slice to volumeregistration
Update slicessegmentation
Robust statisticsand superresolution
Reconstructedvolume
9/12
Motion correction
Conditional Random Field
10/12
Result
Tighter mask, less artifactsand brain already extracted
11/12
Acknowledgements
Collaborators:Maria Murgasova, Vanessa Kyriakopoulou,Christina Malamateniou, Mary Rutherford,Bernhard Kainz, Jo Hajnal & Daniel Rueckert
XKCD for Matplotlib, Latex and Tikz:Jake Vanderplas, Damon McDougall, JohannesBuchner, percusse
www.doc.ic.ac.uk/~kpk09/
12/12
Related Documents
