Projects:

Image reconstruction using Principal component analysis:

In many real life situations, we don’t always have access to perfect datasets. In many industrial systems, the datasets are either sparse or have missing points. Normally, scientists refill these missing points with an estimate, which is usually zeros or the average values. However, principal component analysis provides a expected value which is much more meaningful than zero-filling or using average values. Above algorithm  takes advantage of the colinearity between neighbouring pixels in a 4 dimensional space to provide a best estimate for missing points in a 3 Channel picture.

PCA_PLS_tutorial11

Vessel Tracking using geodesic algorithms:

Geodesic algorithms are the algorithms in GPS systems that find the most optimized path from point A to B. I used one of these algorithms to track blood flow in the brain and also to identify the most efficient pathway for the blood to reach a section of the brain.

Click for additional movies: https://www.youtube.com/edit?o=U&video_id=ypMT_Bohups

 

 

 

 

 

3D rendering of vessels in the neck:

Using a few simple tricks in Matlab, you can create and render very high quality 3D images. you will need Red/Cyan 3D glass in order to view these images:

 

 

 

 

 

Application of PLS-DA method in separation of venous and arterial phase in 4D CE-MRA data:

Contrast enhanced MR venography (CE-MRV) provides useful information about the geometry and flow properties of the vessels. To enhance visualization or assessment, it is often desired to separate the arterial phase from the venous phase.This task can be very complicated especially in patients with venous abnormalities.

Our team developed an algorithm based on a partial least squares discriminant analysis (PLS-DA machine learning) to isolate the venous and arterial phase. More information about the method can be found here:
http://ssalari.net/wordpress/wp-content/uploads/2016/01/separating-venous-and-arterial-phase.pdf

Capture

Application of SC-PLS in removing BCG artifacts in combined EEG-fMRI

An essential step prior to analysis of EEG data in combined EEG-fMRI is to remove the balistocardiographic (BCG) noise. BCG is induced by the body movement or blood circulation when EEG is performed inside the MRI chamber.

We propose a new method based on SC-PLS algorithm to remove the BCG noise which does not require accurate detection of heart rate, and is insensitive to variations in heart rate or motion through the body.

More information about this method can be found here:
click here

Capture3 Capture2