Thesis Project Form

Cochlear implant (CI) is one of the world’s most successful medical implants by restoring the ability to speech understanding in profoundly deaf patients. Cochlear implant surgery involves placing an electrode array inside the scala tympani part of the cochlea. During insertion, the surgeon is blinded by the temporal bone eliminating any chance of visual cues. Sub-optimal insertions often lead to electrode translocation, rupture of basilar membrane, tip-fold overs and electrode buckling, all causing trauma to the patient. Visualizing hidden structures through an augmented reality setup may improve the ergonomics and performance of the procedure.

The aim of this project will be to design an automatic registration strategy between a preoperative CT-scan and real-time video of the temporal bone using visual and optical tracking strategies. The project would involve:
• Point-based initial registration using feedback from optical tracking system.
• Contour based registration optimization.
• Validation on resin phantom models.

The core challenge of the project is the variability of and the low number of distinct features present in the ear anatomy. A hybrid approach will provide the optimal initial registration approach that will be fed into a tracking strategy of synchronizing virtual overlays and patient movements.