Abstract
The cochlear ear implant has become a standard clinical intervention for the treatment of profound sensorineural hearing loss. After 20 years of research into implant design, there are still many unanswered clinical questions that could benefit from new analysis and modelling techniques. This research aims to develop techniques for extracting the cochlea from medical images to support clinical outcomes. We survey the challenges posed by some of these clinical questions and the problems of cochlea modeling. We present a novel algorithm for extracting tubular objects with non-circular cross-sections from medical images, including results from generated and clinical data. We also describe a cochlea model, driven by clinical knowledge and requirements, for representation and analysis. The 3-dimensional cochlea representation described herein is the first to explicitly integrate path and cross-sectional shape, specifically directed at addressing clinical outcomes. The tubular extraction algorithm described is one of very few approaches capable of handling non-circular cross-sections. The clinical results, taken from a human CT scan, show the first extracted centreline path and orthogonal cross-sections for the human cochlea.
| Original language | English |
|---|---|
| Pages (from-to) | 74-85 |
| Number of pages | 12 |
| Journal | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
| Volume | 3339 |
| DOIs | |
| Publication status | Published - 2004 |
| Externally published | Yes |
| Event | 17th Australian Joint Conference on Artificial Intelligence, AI 2004: Advances in Artificial Intelligence - Cairns, Australia Duration: 4 Dec 2004 → 6 Dec 2004 |