Automatic Localization of Curvilinear Object in 3D Ultrasound Images

                         Martin Barva 
  
                    <barvam1@fel.cvut.cz>


Abstract: Utilization of tools during surgical interventions sets the
problem of their accurate localization within biological tissue. Our
task is to determine, in real-time, the position of curvilinear needle
in biological tissue from a three-dimensional ultrasound image.
Initially the data are segmented by thresholding and processed with
the randomized version of the Random Sample Consensus (RANSAC)
algorithm. The curvilinear needle is modeled by a three-dimensional
cubic curve. Its shape is subject to check using a curvature measure
in the hypothesis evaluation step of the R-RANSAC
algorithm. Subsequently, we perform the least squares curve fitting to
the data that have been marked by the R-RANSAC as the ones
corresponding to the sought object. Finally, the localization of the
needle tips is carried out by a hypothesis testing on the distances
between projections of inliers on the estimated curve. From the
results, we conclude that the method is very stable even if the data
contain high percentage of outliers. The error of position estimation
is minimal with respect to the mean square criterion. The
computational cost of the algorithm shows the possibility of real-time
data processing.