Automatic Cochlear Length and Volume Size Estimation

摘要

Surgical robotics tools using real-time computer vision algorithms can detect, extract and analyze specific organs during surgery. In order to achieve successful cochlear implant surgical operation using such techniques, the exact cochlear length and size are required. We present a fast cochlear length and volume size estimation method that may be integrated in such situations. The method utilizes atlas-model-based segmentation to estimate a transformation from a model to an input volume. The result is used to transform a well-defined segmentation and a points-set of a scala tympani to the input image that segments and estimates the scala tympani length in a few seconds. Given the lack of a publicly-available ground truth, the error is estimated using the known length of the cochlear implants. A dataset of 71 3D images of 21 patients from various age and gender groups is used. The estimated average scala tympani length is 29.54 mm, with 0.27 mm standard deviation. The average scala tympani volume size is 41.56 mm~3, with 0.19 mm~3 standard deviation. The method is available as an open-source 3D Slicer plug-in and the dataset is also publicly-available.