Nobuo Suzumura

Learn More
A data-compression algorithm for digital Holter recording using artificial neural networks (ANNs) is described. A three-layer ANN that has a hidden layer with a few units is used to extract features of the ECG (electrocardiogram) waveform as a function of the activation levels of the hidden layer units. The number of output and input units is the same. The(More)
Conventional X-ray tomosynthesis with film can provide a sagittal slice image with a single scan. This technique has the advantage of enabling reconstruction of a sagittal slice which is difficult to obtain from the X-ray CT system. However, only an image on the focal plane is obtained by a single scan. Furthermore, the image is degraded by superimpositions(More)
A new automatic computer program, to remove misregistration artifacts secondary to motions such as rotation, contraction and relaxation of the body, has been developed. This program is called automatic reregistration and consists of determination of distortion vectors of corresponding segments of the mask and live images by detection of the peak value of(More)
In this paper, we show techniques to examine the stationarity and the normality of time series as well as results obtained by applying these techniques to EEG data during sleep stages. Many statistical analyses of the EEG data are based on the assumption that the EEG data are stationary and normally distributed. However, the problem is to know the length of(More)
Biomedical data, such as EEG, EMG and neural impulse sequences, are regarded as the stochastic phenomena of biological systems, and the statistical properties of such time series are often examined. Most of the statistical analysis processed in the frequency and the time domain are based on the assumption that the time series is weakly stationary and(More)
A detection algorithm for the first and the second heart sounds, which is one of the most important problems in an automatic diagnostic system for phonocardiograms, has been developed. It is based on the frequency-domain characteristics of heart sounds analysed by a linear-prediction method. The performance of the algorithm has been evaluated in 187 samples(More)
The authors investigated the usefulness of the Karhunen-Loeve technique applied to body surface maps to study regional cardiac excitation. Eigenvectors were derived from the body surface potential maps of 120 healthy adults using the Karhunen-Loeve expansion theory. Then, in the maps of various types of ventricular hypertrophy, each eigenvector coefficient(More)
The paper reports some results with phonocardiogram pattern classification. Linear prediction analysis was applied to extract the spectral pattern from phonocardiogram signals, a relatively new technique. In this examination, 29 design samples are classified correctly, except for three samples; and 19 test samples correctly, except for five samples. The(More)
A data compression algorithm for digital Holter recording using artificial neural networks (ANN) is proposed. A dual three-layer (one hidden layer) neural network which has a few units of hidden layer is used to extract the differences of waveforms as the activation levels of hidden layer units. The network is tuned using supervised signals, which are the(More)