Biomedical subjects

Cishen Zhang

Publications and source records attributed to Cishen Zhang.

4 recordsLinked to original sources

PubMed2006

New necessary and sufficient conditions for absolute stability of neural networks.

This paper presents new necessary and sufficient conditions for absolute stability of asymmetric neural networks. The main result is based on a solvable Lie algebra condition, which generalizes existing results for symmetric and normal neural networks. An exponential convergence estimate of the neural networks is also obtained. Further, it is demonstrated how to generate larger sets of weight matrices for absolute stability of the neural networks from known normal weight matrices through simple procedures. The approach is nontrivial in the sense that non-normal matrices can possibly be contained in the resulting weight matrix set. And the results also provide finite checking for robust stability of neural networks in the presence of parameter uncertainties.

Algorithms↗

PubMed2006

Constrained least squares filtering algorithm for ultrasound image deconvolution.

A new medical ultrasound tissue model is considered in this paper, which incorporates random fluctuations of the tissue response and provides more realistic interpretation of the received pulse-echo ultrasound signal. Using this new model, we propose an algorithm for restoration of the degraded ultrasound image. The proposed deconvolution is a modification of the classical regularization technique which combines Wiener filter and the constrained least squares (LS) algorithm for restoration of the ultrasound image. The performance of the algorithm is evaluated based on both the simulated phantom images and real ultrasound radio frequency (RF) data. The results show that the algorithm can provide improved ultrasound imaging performance in terms of the resolution gain. The deconvolved images visually show better resolved tissue structures and reduce speckle, which are confirmed by a medical expert.

Algorithms↗

PubMed2005

Tilted plane Feldkamp type reconstruction algorithm for spiral cone beam CT.

An approximate image reconstruction method for spiral cone beam computed tomography (CT), called tilted plane Feldkamp type reconstruction algorithm (TPFR), is presented in this paper, which extends Feldkamp cone beam reconstruction algorithm to deal with its inaccuracy and artifact problems caused by large cone angle. This is done by tilting the reconstructing planes to minimize the cone angle and optimally fit the spiral segment of the source. The tilted plane image reconstruction requires reforming the three-dimensional projection data set for the tilted plane and application of Feldkamp algorithm to the reformed data set. Analytical and computational results can show that the image reconstruction performance of the proposed TPFR algorithm is superior to that of the Feldkamp reconstruction algorithm in the image quality, volume coverage speed, maximum achievable pitch value, and slice sensitivity profiles. Moreover, it provides more accurate image reconstruction than the existing two-dimensional reconstruction algorithms.

Algorithms↗

PubMed2003

Necessary and sufficient condition for absolute stability of normal neural networks.

Globally convergent dynamics of a class of neural networks with normal connection matrices is studied by using the Lyapunov function method and spectral analysis of the connection matrices. It is shown that the networks are absolutely stable if and only if all the real parts of the eigenvalues of the connection matrices are nonpositive. This extends an existing result on symmetric neural networks to a larger class including certain asymmetric networks. Further extension of the present result to certain non-normal case leads naturally to a quasi-normal matrix condition, which may be interpreted as a generalization of the so-called principle of detailed balance for the connection weights or the quasi-symmetry condition that was previously proposed in the literature in association with symmetric neural networks. These results are of particular interest in neural optimization and classification problems.

Models, Theoretical↗