A video about Fast Fully Automatic Brain Detection in Fetal MRI Using Dense Rotation Invariant Image Descriptors can be found here:
I'll give some updates about the dataset paper in a later post.
Dienstag, 3. Dezember 2013
2 ISBI paper accepted!!
We have successfully submitted a dataset paper about the porcine study we have done in the course of the IMPPACT project in Graz and a paper about automatic brain localization and segmentation in fetal MRI to ISBI 2014.
A video about Fast Fully Automatic Brain Detection in Fetal MRI Using Dense Rotation Invariant Image Descriptors can be found here:
I'll give some updates about the dataset paper in a later post.
A video about Fast Fully Automatic Brain Detection in Fetal MRI Using Dense Rotation Invariant Image Descriptors can be found here:
I'll give some updates about the dataset paper in a later post.
Donnerstag, 28. November 2013
Samstag, 16. November 2013
530 - Software Engineering Practice and Group project available
The MRI foetus detector
Recently, several novel MRI sequences have been developed to scan a foetuses inside the womb. The resulting 3D scans show, besides the foetus, a significant amount of maternal tissue and are also subject to movement artefacts caused by the foetus. An automatic evaluation of the foetal organs would be desirable but is currently difficult because of the large amount of background information. The aim of this project is to suppress the maternal background information and to easy subsequent processing of pre-natal foetal MRI datasets.
The key objectives of this project are therefore:
- to implement state-of-the-art object detection algorithms, using existing libraries, and to train and evaluate their performance on the individual slices of the 3D datasets.
- to implement an object detector for 3D tubular structures to find the spine of the foetus directly in the volumes and to compare this approach to the slice based approch above.
- to use machine learning for an automatically generated probability map and to visualize the likelihood of foetal tissue.
The project should be implemented in Matlab or C/C++ running on a Desktop PC. Excellent programming skills and experience in image processing and machine learning are desirable.
Recently, several novel MRI sequences have been developed to scan a foetuses inside the womb. The resulting 3D scans show, besides the foetus, a significant amount of maternal tissue and are also subject to movement artefacts caused by the foetus. An automatic evaluation of the foetal organs would be desirable but is currently difficult because of the large amount of background information. The aim of this project is to suppress the maternal background information and to easy subsequent processing of pre-natal foetal MRI datasets.
The key objectives of this project are therefore:
- to implement state-of-the-art object detection algorithms, using existing libraries, and to train and evaluate their performance on the individual slices of the 3D datasets.
- to implement an object detector for 3D tubular structures to find the spine of the foetus directly in the volumes and to compare this approach to the slice based approch above.
- to use machine learning for an automatically generated probability map and to visualize the likelihood of foetal tissue.
The project should be implemented in Matlab or C/C++ running on a Desktop PC. Excellent programming skills and experience in image processing and machine learning are desirable.
Mittwoch, 13. November 2013
money rules
Today I got the official confirmation at a PostDoc Development Centre course that basically the only criteria to stay in academia is how much money you bring the university (grants etc). It is actually absolutely not evaluated how good you're in teaching. I would understand this attitude when it comes to Austrian universities, which don't take study fees. However, in UK the study fees are high and still they do not aim for high quality teachers. They only aim for cash cows. Why do the universities all over the world want to follow this system? This is comparable to hiring a bus driver according to her/his ability to motivate customers to get on the bus without checking her/his driving history or driving license. Anyway, if money and prestige are really the only things that matter in academia, then shouldn't they outsource the teaching part completely?
btw: I'll soon post research related stuff again :)
Sonntag, 8. September 2013
New student projects available
I'm offering a few very interesting student projects (MSc, Mres) this
year for students of Imperial College London. If you're interested log
in at https://cate.doc.ic.ac.uk and select one of my projects:
------------------------------------------------------------------------------------------
Machine learning for the automatic evaluation of foetal movements in-utero
Currently, the face of prenatal diagnostics is changing rapidly. Novel MRI sequences allow to make videos of the foetal movements inside the womb with a large field of view. Thereby, the assessment of motor function is an essential component of neurology examinations. Recent research shows evidence that it is possible to assess the healthiness of a foetus by observing the kind and extend of movements. An automatic pre-classification in healthy and abnormal foetal behaviour would form a valuable tool for the clinical practise.
The key objectives of this project are therefore:
- to implement state-of-the-art motion estimation algorithms, such as optical flow, using existing libraries and to compare their performance.
- to derive significant features from the motion analysis and to use machine learning techniques to distinguish between healthy and abnormal foetal behaviour.
Bonus objective:
The datasets show a combination of maternal movements and foetal movements, which are not directly related. Therefore, a desirable additional objective would be to extract movement features that are only related to the foetus.
The project should be implemented in Matlab or C/C++ running on a Desktop PC. Excellent programming skills and experience in image processing and machine learning are desirable.
------------------------------------------------------------------------------------------
The MRI foetus detector
Recently, several novel MRI sequences have been developed to scan a foetuses inside the womb. The resulting 3D scans show, besides the foetus, a significant amount of maternal tissue and are also subject to movement artefacts caused by the foetus. An automatic evaluation of the foetal organs would be desirable but is currently difficult because of the large amount of background information. The aim of this project is to suppress the maternal background information and to easy subsequent processing of pre-natal foetal MRI datasets.
The key objectives of this project are therefore:
- to implement state-of-the-art object detection algorithms, using existing libraries, and to train and evaluate their performance on the individual slices of the 3D datasets.
- to use machine learning for an automatically generated probability map, showing the likelihood of foetal tissue.
The project should be implemented in Matlab or C/C++ running on a Desktop PC. Excellent programming skills and experience in image processing and machine learning are desirable.
------------------------------------------------------------------------------------------
Uncertainty visualization of foetal brain super-resolution reconstructions
Super-resolution reconstruction of the foetal brain got recently possible through the development of novel MRI sequences for the womb. Thereby, several low resolution image stacks of the foetus are acquired in overlapping but different directions. These image stacks can subsequently be used to reconstruct a high resolution volume of the desired region, usually the foetal brain. However, the resulting images show errors that are not directly noticeable. Therefore, a useful clinical extension of this approach would be to develop a tool that is able to visualize the errors, which occur during the reconstruction process.
The key objectives of this project are therefore:
- to extend an existing framework for the reconstruction of foetal MRI data, so that additional information about the uncertainty of the calculated intensity values are generated.
- to develop a tool for the visualization of these errors. Thereby, the uncertainty information should not distract the user from the anatomical information and state-of-the-art uncertainty visualization methods should be used.
The project should be implemented in C/C++ running on a Desktop PC. The visualization can be integrated into existing visualization software. Excellent programming skills and experience in image processing and machine learning are desirable.
------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------
Machine learning for the automatic evaluation of foetal movements in-utero
Currently, the face of prenatal diagnostics is changing rapidly. Novel MRI sequences allow to make videos of the foetal movements inside the womb with a large field of view. Thereby, the assessment of motor function is an essential component of neurology examinations. Recent research shows evidence that it is possible to assess the healthiness of a foetus by observing the kind and extend of movements. An automatic pre-classification in healthy and abnormal foetal behaviour would form a valuable tool for the clinical practise.
The key objectives of this project are therefore:
- to implement state-of-the-art motion estimation algorithms, such as optical flow, using existing libraries and to compare their performance.
- to derive significant features from the motion analysis and to use machine learning techniques to distinguish between healthy and abnormal foetal behaviour.
Bonus objective:
The datasets show a combination of maternal movements and foetal movements, which are not directly related. Therefore, a desirable additional objective would be to extract movement features that are only related to the foetus.
The project should be implemented in Matlab or C/C++ running on a Desktop PC. Excellent programming skills and experience in image processing and machine learning are desirable.
------------------------------------------------------------------------------------------
The MRI foetus detector
Recently, several novel MRI sequences have been developed to scan a foetuses inside the womb. The resulting 3D scans show, besides the foetus, a significant amount of maternal tissue and are also subject to movement artefacts caused by the foetus. An automatic evaluation of the foetal organs would be desirable but is currently difficult because of the large amount of background information. The aim of this project is to suppress the maternal background information and to easy subsequent processing of pre-natal foetal MRI datasets.
The key objectives of this project are therefore:
- to implement state-of-the-art object detection algorithms, using existing libraries, and to train and evaluate their performance on the individual slices of the 3D datasets.
- to use machine learning for an automatically generated probability map, showing the likelihood of foetal tissue.
The project should be implemented in Matlab or C/C++ running on a Desktop PC. Excellent programming skills and experience in image processing and machine learning are desirable.
------------------------------------------------------------------------------------------
Uncertainty visualization of foetal brain super-resolution reconstructions
Super-resolution reconstruction of the foetal brain got recently possible through the development of novel MRI sequences for the womb. Thereby, several low resolution image stacks of the foetus are acquired in overlapping but different directions. These image stacks can subsequently be used to reconstruct a high resolution volume of the desired region, usually the foetal brain. However, the resulting images show errors that are not directly noticeable. Therefore, a useful clinical extension of this approach would be to develop a tool that is able to visualize the errors, which occur during the reconstruction process.
The key objectives of this project are therefore:
- to extend an existing framework for the reconstruction of foetal MRI data, so that additional information about the uncertainty of the calculated intensity values are generated.
- to develop a tool for the visualization of these errors. Thereby, the uncertainty information should not distract the user from the anatomical information and state-of-the-art uncertainty visualization methods should be used.
The project should be implemented in C/C++ running on a Desktop PC. The visualization can be integrated into existing visualization software. Excellent programming skills and experience in image processing and machine learning are desirable.
------------------------------------------------------------------------------------------
Montag, 2. September 2013
Freitag, 12. Juli 2013
IEEE SciVis 2013 paper accepted
Rostislav's, Markus's, Dieter's and my paper on "Noise-based volume rendering for the visualization of multivariate volumetric data" got accepted for IEEE SciVis 2013!
The paper is about using artificial but ingenious noise patterns for volume rendering to communicate hidden and multi-dimensional information of the data. Besides the scientific value of this new approach, the resulting pictures are beautiful!
The paper is about using artificial but ingenious noise patterns for volume rendering to communicate hidden and multi-dimensional information of the data. Besides the scientific value of this new approach, the resulting pictures are beautiful!
 |
| A part of a noise enhanced visualization of multi-dimensional weather data. |
Donnerstag, 11. Juli 2013
Updated profile on research gate
After a very interesting workshop from Imperial's postdoc development centre, I've deceided to update my ResearchGate Profile. Have a look here: https://www.researchgate.net/profile/Bernhard_Kainz2
Montag, 6. Mai 2013
Project started!!
So, we did it! I got the contract for the upcoming two years and can start to work on my own research project at Imperial College London. Thanks' to the European Commission (EC) the signed grant agreement arrived also in time.
The image shows one slice from a 3D MRI volume scan of a pregnant women. This is one slice of one volume out of around 200 datasets that are available for F.A.U.S.T. I'm looking forward to attend the upcoming scan sessions to acquire even more datasets. |
 |
| One slice out of many many datasets... |
Dienstag, 2. April 2013
Ethical screening successfully done - project start on 01.05.2013
This blog is
part of my Marie Curie Fellowship project: F.A.U.S.T. -- Flexible Application
of Uncertainty for Scanning and Tracking. This project is currently in its
Grant Agreement phase and will start on 01.05.2013.
F.A.U.S.T
will deal with reconstruction of fetal Magnetic Resonance Imaging (MRI) data
that are taken directly from the female uterus. MRI is a harmless imaging
procedure, which does not use any radiation or other invasive procedures to get
the images. However, the fetus is awake and will move around during the images
are taken. This is where my colleges and I come into play. We try to
reconstruct the fetus in 3D as if it would not move. Furthermore, I'll try to
evaluate these steps for their reliability and I will search for methods that
allow a doctor to draw conclusions on the reliability of each part of the
resulting images.
The data acquired
at St. Thomas Hospital London by King's College London and is processed at the
Biomedical Image Processing Group at Imperial College London, where I work together with Prof. Daniel Rueckert. My own background is currently still summarized best on my staff page from my former Institute for Computer Graphics and Vision at Graz University of Technology.
I'll try to
keep the general public and scientific community up to date with this blog
about
- software developments
- publications
- videos and results about the reconstruction process
- research flops (if a certain direction is not successful ;) )
I'm looking
forward to get many interested readers!
Abonnieren
Posts (Atom)