Conclusion
At the finale of this project I have investigated the properties of
filtering tools on fMRI data.
I have been able to
extend existing work on current filters and develop new filters that can
be used on fMRI.
I have acheived this by using existing work as a base and modify it to use
different techniques to possibly achieve better results.
The tests carried out were on data from a well recognised group involved
in fMRI, namely FMRIB.
The results on images with artificial noise turned out better for the
Vector Median Filter but for images that were free from artificial noise,
the Least Change Filters were the preferred choice.
This project could continue by filtering in dimensional modes higher than
2.
In addition, it would be useful to see how voxel values in individual
volumes vary over time.