Abstract
A novel approach of applying artificial neural network techniques to three-dimensional (3-D) nonrigid motion analysis is proposed. The research presented in this paper examines the 3-D nonrigid motion of the left ventricle of a human heart through use of biplanar cineangiography data, consisting of 3-D coordinates of 30 coronary artery bifurcation points of the left ventricle and the correspondences of these points taken over 10 time instants during the heart cardiac cycle. The 3-D nonrigid motion of the left ventricle of a heart can be decomposed into both a global rigid motion and a set of local nonrigid deformations, which are coupled with the global motion at every time instant [1]-[3]. Using the artificial neural network techniques proposed in [4], the global rigid motion can be estimated precisely in the mathematic form of a translation vector and a rotation matrix. Estimation of the local nonrigid deformations will be discussed in this paper. A set of neural networks similar in structure and dynamics but different in physical size is proposed to tackle the problem of nonrigidity in the local deformation estimation. These neural networks are parallel in operation and connected to each other through feedbacks. Each network consists of an input layer and an output layer. The activation function of the output layer is selected in such a way that a feedback is involved in the output updating. The constraints are specified to ensure a stable and globally consistent estimation of local deformations. The constraints are incorporated into the network structure through the assignment of weights between two layers, the initial values of the outputs, as well as the connections between each network. The objective of the proposed neural networks is to find the optimal deformation matrices that satisfy the constraints for all the coronary artery bifurcation points of the left ventricle. The proposed neural networks differ from other existing neural network models in their unique structure and dynamics. Experiments on the biplanar cineangiography data are conducted to corroborate the proposed techniques.
Original language | English (US) |
---|---|
Pages (from-to) | 1394-1401 |
Number of pages | 8 |
Journal | IEEE Transactions on Neural Networks |
Volume | 6 |
Issue number | 6 |
DOIs | |
State | Published - Nov 1995 |
ASJC Scopus subject areas
- Software
- Computer Science Applications
- Computer Networks and Communications
- Artificial Intelligence