An Interacting Multiple Model Probabilistic Data Association Filter for Cavity Boundary Extraction From Ultrasound Images

Abstract

This paper presents a novel segmentation technique for extracting cavity contours from ultrasound images. The problem is first discretized by projecting equispaced radii from an arbitrary seed point inside the cavity toward its boundary. The distance of the cavity boundary from the seed point is modeled by the trajectory of a moving object. The motion of this moving object is assumed to be governed by a finite set of dynamical models subject to uncertainty. Candidate edge points obtained along each radius include the measurement of the object position and some false returns. The modeling approach enables us to use the interacting multiple model estimator along with a probabilistic data association filter, for contour extraction. The convergence rate of the method is very fast because it does not employ any numerical optimization. The robustness and accuracy of the method are demonstrated by segmenting contours from a series of ultrasound images. The results are validated through comparison with manual segmentations performed by an expert. An application of the method in segmenting bone contours from computed tomography images is also presented.

3D Prostate Segmentation in Ultrasound Images Based on Tapered and Deformed Ellipsoids

“…The resulting tapered ellipse is used to guide the IMMPDA edge detection algorithm [17] by setting limits on how far from the initial contour the edge detection can search. In order to improve the tapered ellipse fitting, the resulting edge points are once again fed to the LevenbergMarquardt algorithm.…”

Estimating 3D lumen centerlines of carotid arteries in free-hand acquisition ultrasound

“…The publications most related to our research are from Abolmaesumi and Sirouspour [13], Golemati et al [14], Hammer et al [15], and Wang et al [16,17]. Abolmaesumi & Sirouspour developed a technique to extract boundaries for cavities in ultrasound images.…”

Contour segmentation in 2D ultrasound medical images with particle filtering

“…Tracking methods are adopted also in [1] for the purposes of heart chambers and breast cyst segmentation. The recursive contour growing (motion) is governed by a finite set of switching dynamical models and thus its behavior can be probabilistically predicted.…”

“…The concept of combining multiple trajectory models in order to estimate the state of manoeuvring object in clutter has its rational solution in the face of combining the interacting multiple model (IMM) estimator and a probabilistic data association filter (PDAF) [6]. The authors of [1,2] demonstrate the accuracy of their IMM-PDAF implementation by segmenting convex, non-circular lesion forms over a number of prostate, carotid artery, jugular vein ultrasound images. The IMM-PDAF results show that the multiple hypothesis approach jointly applied with multiple switching motion models yields correct and convergent contour tracking in the complicated medical imaging environment.…”

“…The positions of four manually selected points are required: a seed point, a starting point and two additional points, bounding the measurement formation area around the contour. In some medical applications, where additional information is available, the proposed approach can make use of two points only and hence to reduce in this way the operator intervention, similarly to [1,2].…”