Registration of micro-computed tomography and histological images of the guinea pig cochlea to construct an ear model using an iterative closest point algorithm.

摘要

We present a practical and systematic method to reconstruct accurate physical models of the guinea pig ear (n = 1). The method uses a semi-automatic technique to create three-dimensional (3-D) models of the guinea pig cochlea by registration of micro-computed tomography (CT) and histological images. An iterative closest point algorithm was employed to minimize the sum of square errors with respect to the closest histological model and corresponding micro-CT model. This allowed creation of an accurate geometric ear model including external ear canal, tympanic membrane, middle ear cavity, auditory ossicles, and the cochlea. The characteristic cross-sectional areas of scala tympani, scala vestibuli, and scala media were measured. The length, thickness, and apex width of the guinea pig's basilar membrane were compared to the data found in literature. Some shape parameters were also compared among different species. The results confirmed that the geometric model created by this method was accurate. This method provides an effective way to visualize the 3-D structure and the detailed information about ear geometry required for finite element and multibody dynamic analysis.