Greg D. Parker1,
2, David Marshall2, Paul L. Rosin2, Nicholas
Drage3, Stephen Richmond3, Derek K. Jones1
1CUBRIC,
School of Psychology, Cardiff University, Cardiff, South Glamorgan, United
Kingdom; 2School of Computer Science, Cardiff University, Cardiff,
South Glamorgan, United Kingdom; 3School of Dentistry, Cardiff University,
Cardiff, South Glamorgan, United Kingdom
We present a novel method for robust estimation of fibre orientation distribution functions through diffusion weighted signal outlier rejection. Our algorithm combines aspects of the Richardson-Lucy spherical deconvolution with a non negative sparse coder to produce an adaptive signal dictionary that 'learns' compensations for common non-axially symmetric noise/corruption while preserving signals arising from complex fibre architecture. This has the effect of improving both fODF estimation and (by studying the dictionary adaptations) the robustness of outlier identification.