@MASTERSTHESIS\{IMM2007-05117,
    author       = "F. S. Telcean",
    title        = "Structure from Motion Methods for {3D} Modeling of the Ear Canal",
    year         = "2007",
    school       = "Informatics and Mathematical Modelling, Technical University of Denmark, {DTU}",
    address      = "Richard Petersens Plads, Building 321, {DK-}2800 Kgs. Lyngby",
    type         = "",
    note         = "Supervised by Bjarne Kj{\ae}r Ersb{\o}ll and co-supervised by Henrik Aan{\ae}s, {IMM,} {DTU}.",
    url          = "http://www2.compute.dtu.dk/pubdb/pubs/5117-full.html",
    abstract     = "Structure from Motion deals with {3D} reconstruction from {2D} images and it is one of the most widely researched problems within computer vision area. Recently, it was successfully integrated in many medical oriented applications.

Reconstruction of accurate models of the ear canal is a key step in the design and production of hearing aids. Actual methods are based on an invasive procedure (ear canal impression taking), require time, special trained skills and sophisticated and expensive hardware as {3D} laser scanners. On the other side, the video otoscope became a standard tool in the hearing specialist office and it is able to provide images of the ear canal.

This thesis is about {3D} reconstruction of the human ear canal from images using Structure from Motion methods. Two aspects are studied. First, the images of the ear canal are analyzed in order to see if they provide enough information for reconstruction algorithms. Second, the reconstruction accuracy of tube-like objects is analyzed in the context of a specific Structure from Motion algorithm."
}