Determination and quantitative evaluation of image-based registration accuracy for robotic neurosurgery

Cutter, Jennifer Ruth (2018). Determination and quantitative evaluation of image-based registration accuracy for robotic neurosurgery. University of Birmingham. Ph.D.

Preview

Cutter18PhD.pdf
PDF - Accepted Version
Download (7MB)

Abstract

Stereotactic neurosurgical robots allow quick, accurate location of small targets within the brain, relying on accurate registration of preoperative MRI/CT images with patient and robot coordinate systems. Fiducial markers or a stereotactic frame are used as registration landmarks and the patient’s head is fixed in position. An image-based system could be quick, non-invasive and allow the head to be moved during surgery giving greater ease of access. Submillimetre surgical precision at the target point is required.
An octant representation is utilized to investigate full region of interest (ROI) head registration using parts only, with registration performed using the Iterative Closest Point (ICP) algorithm. Use of two octants sequentially obtained a mean RMS distance of 0.813±0.026 mm; adding subsequent octants did not significantly improve performance. An RMS distance of 0.812±0.025 mm was obtained for three octants used simultaneously.
ICP was compared with Coherent Point Drift, and 3D Normal Distribution Transform, with and without added or smoothed noise, and was least affected by starting position or noise added; a mean accuracy of 0.884±0.050 mm across ten noise levels and four starting positions was achieved, which was shown to translate to submillimetre accuracy at points within the head.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):