Biomedical subjects

G Serpen

Publications and source records attributed to G Serpen.

2 recordsLinked to original sources

PubMed2003

Automated lung outline reconstruction in ventilation-perfusion scans using principal component analysis techniques.

The present work addresses the development of an automated software-based system utilized in order to create an outline reconstruction of lung images from ventilation-perfusion scans for the purpose of diagnosing pulmonary embolism. The proposed diagnostic software procedure would require a standard set of digitized ventilation-perfusion scans in addition to correlated chest X-rays as key components in the identification of an ideal template match used to approximate and reconstruct the outline of the lungs. These reconstructed lung images would then be used to extract the necessary PIOPED-compliant features which would warrant a pulmonary embolism diagnosis. In order to evaluate this issue, two separate principal component analysis (PCA) algorithms were employed independently, including Eigenlungs, which was adapted from the Eigenfaces method, and an artificial neural network. The results obtained through MATLAB(TM) simulation indicated that lung outline reconstruction through the PCA approach carries significant viability.

Algorithms↗

PubMed2001

The simultaneous recurrent neural network for addressing the scaling problem in static optimization.

A trainable recurrent neural network, Simultaneous Recurrent Neural network, is proposed to address the scaling problem faced by neural network algorithms in static optimization. The proposed algorithm derives its computational power to address the scaling problem through its ability to "learn" compared to existing recurrent neural algorithms, which are not trainable. Recurrent backpropagation algorithm is employed to train the recurrent, relaxation-based neural network in order to associate fixed points of the network dynamics with locally optimal solutions of the static optimization problems. Performance of the algorithm is tested on the NP-hard Traveling Salesman Problem in the range of 100 to 600 cities. Simulation results indicate that the proposed algorithm is able to consistently locate high-quality solutions for all problem sizes tested. In other words, the proposed algorithm scales demonstrably well with the problem size with respect to quality of solutions and at the expense of increased computational cost for large problem sizes.

Algorithms↗