3-D Heart Modeling and Motion EstimationBased on Continuous Distance Transform Neural Networks andAffine Transform

Authors:
Yen-Hao Tseng;Jenq-Neng Hwang;Florence H. Sheehan
Affiliations:
Information Processing Laboratory, Dept. of Elect. Engr., Box 352500, University of Washington, Seattle, WA 98195, USA;Information Processing Laboratory, Dept. of Elect. Engr., Box 352500, University of Washington, Seattle, WA 98195, USA;Division of Cardiology, School of Medicine, University of Washington, Seattle, WA 98195, USA
Venue:
Journal of VLSI Signal Processing Systems - special issue on applications of neural networks in biomedical image processing
Year:
1998

Citing 12
Cited 2

Perceptual organization and the representation of natural form

Artificial Intelligence
The representation, recognition, and locating of 3-d objects

International Journal of Robotics Research
On Optimally Combining Pieces of Information, with Application to Estimating 3-D Complex-Object Position from Range Data

IEEE Transactions on Pattern Analysis and Machine Intelligence
Hierarchical Chamfer Matching: A Parametric Edge Matching Algorithm

IEEE Transactions on Pattern Analysis and Machine Intelligence
Multilayer feedforward networks are universal approximators

Neural Networks
On Three-Dimensional Surface Reconstruction Methods

IEEE Transactions on Pattern Analysis and Machine Intelligence
Estimation of Planar Curves, Surfaces, and Nonplanar Space Curves Defined by Implicit Equations with Applications to Edge and Range Image Segmentation

IEEE Transactions on Pattern Analysis and Machine Intelligence
Describing Complicated Objects by Implicit Polynomials

IEEE Transactions on Pattern Analysis and Machine Intelligence
The Complex EGI: A New Representation for 3-D Pose Determination

IEEE Transactions on Pattern Analysis and Machine Intelligence
Parameterized Families of Polynomials for Bounded Algebraic Curve and Surface Fitting

IEEE Transactions on Pattern Analysis and Machine Intelligence
Information extraction about complex three-dimensional objects from visual data (bayesian, position estimation, robot vision, recognition)

Information extraction about complex three-dimensional objects from visual data (bayesian, position estimation, robot vision, recognition)
A neural network-based stochastic active contour model (NNS-SNAKE) for contour finding of distinct features

IEEE Transactions on Image Processing

A note on future research in segmentation techniques applied to neurology, cardiology, mammography and pathology

Advanced algorithmic approaches to medical image segmentation
2D Euclidean distance transform algorithms: A comparative survey

ACM Computing Surveys (CSUR)

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, we apply the previously proposedcontinuous distance transform neural network (CDTNN) to represent3-D endocardial (inner) and epicardial (outer) contours andquantitatively estimate the motion of left ventriclesof human hearts from ultrasound images acquired usingtransesophageal echo-cardiography. This CDTNN has many goodproperties as the conventional distance transforms, which aresuitable for 3-D object representation and deformation estimation. We have successfully represented the 3-D epicardia and endocardiaof left ventricles using CDTNNs trained by as few as 7.5% of themanually traced data. The mean absolute error in the testing forone patient over the 27 testing planes were (1.4 ± 1.2 mm)for the endocardium, (1.3 ± 1.0 mm) for the epicardium at end diastole and (1.4 ± 1.2 mm) for theendocardium vs. 1.2 ± 1.0 mm for the epicardium at endsystole. The absolute error measured compares favorably with thehuman inter-observer variability reported for analyzing distances.With this unique distance transform representation that iscontinuous and differentiable, we are also able to systematicallyand effectively measure the amount of 3-D heart motion in terms ofaffine transform.