60 likes | 182 Vues
Dr. Keith Evan Schubert presents innovative MRI correction methods that enhance target localization in functional neurosurgery for conditions like Parkinson’s disease and trigeminal neuralgia. By employing advanced image processing techniques, the correction process achieves an error reduction of less than 1 mm, transforming over 250,000 pixels per study. This methodology utilizes spherical harmonics to distinguish tissue types and applies ideal image transformations, effectively minimizing gradient nonlinearity distortion. The results meet essential medical accuracy requirements, ensuring more precise treatment outcomes.
E N D
MRI CorrectionIn 5-7 Minutes Dr. Keith Evan Schubert
Situation • Functional Neurosurgery • Parkinson’s Disease, Trigeminal Neuralgia • Error < 1mm • MRI target localization • Distinguish tissue types • More than 250k pixels/study • Gradient Nonlinearity Distortion ~2mm error
Basic Process Select Slices Find Edges Fit Spherical Harmonics Ideal Image Transform Apply
Spherical Harmonics • Cylinder’s Magnetic field • (x2+y2), z2 • Combinations & multiples • Pixel locations → b • Basis functions & Ideal → A • Unknown Coefficients → x • Solve Ax=b for x
Corrected Images Coronal Axial Sagital
Conclusions • Error reduction • Was 2mm or more • 5 to 8 pixels • σ≈ .15mm to .25mm • Less than 1 pixel • 3σ < .8mm • About 2 pixels • Meets medical needs