Automatic classification of sleep staging by electroencephalogram (EEG) is one of the problems that has remained in the field of biomedical engineering. The medical doctor or the qualified electroencephalographer takes much time and work to execute the sleep staging by the recorded papers of the sleeping EEG.
In this study, a method for automatic classification of the sleep staging based on the power spectrum features of the EEG, and on the EOG and the EMG has been developed. First, the features of the EEG on the frequency domain are extracted by applying the autoregressive model to the sleeping EEG. The features are characterized in terms of the coefficients of the autoregressive model. Then distances of a maximum likelihood statistic are computed by using the coefficients.Thus the characteristic waves are determined on the sleeping EEG in due order. Next, from the characteristic waves and information on the eye movement as well as on the activity of the EMG, the sleep staging is carried out. This staging is based on the sleep stage manual by Rechtschaffen and Kales. Finally, these results calculated by the computer are compared with those of a medical doctor. These techniques developed here will be applicable to the anesthesia level EEG classification.

抄録全体を表示

A photo-coupled biomeasuring system using a liquid crystal optical modulator has been studied experimentally. Using this system, consisting of an optical transmitter, a receiver and two optical fibers, patients are fully isolated from ac lines. The system helps to make patient monitoring safe as electric shock hazard have been eliminated. Excessive noise induction along the metallic leads can also be avoided. Practical use of photo-coupled biomeasuring system primarily depends upon the power consumption of the optical transmitter energized by dry batteries. A liquid crystal optical modulator is introduced as the driving power of optical modulator is fairly lower than driving LED's. The optical modulator made of the conventional liquid crystal display panel material usually exhibits very slow responses. However, if it is driven by AM signals, the optical output becomes linear to the envelope of the AM signals, and its frequency response extends up to a fairly high level of eq. 500 Hz, with-6dB/oct drop. The amplitudes of optical signals are distorted by motion of the fibers. In optical FM signals, this distortion is avoided, because the amplitudes of FM signals have no information. Then, the liquid crystal optical modulator is driven by FM-AM technique to make optical output signals into FM waveforms. It can be realized that an overall frequency response of the system is fiat between 0.1 Hz and 100 Hz with the power consumption of 1 mW in the optical transmitter. The characteristics of the system is verified by recording ECG's generated by an ECG generator.

抄録全体を表示

This paper discusses the problem of the feature extraction from the standing position-anteroposterior radiogram on the statistical basis as a part of the study for the purpose of computerizing the mass screening of stomachs with a particular interest in the deformity of the apex, or the sulcus angularis.
As to a number of medically confirmed samples, features were evaluated on their performance to discriminate among the normal samples, ulcerated samples and cancerous samples by a typical method of the discriminant analysis, the minimum Mahalanobis' distance classification. The error rate was estimated by the leaving-one-out method and then the best sets of discriminative features were selected on the basis of the forward sequential procedure. The separability of the samples and the characteristics of the selected features were also inspected by the cluster analysis.
Two approaches were examined in the feature extraction. One consisted of extraction from the contour of the stomach. The best sets of features were selected from among five geometrical measurements and twenty-eight Fourier descriptors. In the best N/A, or normal/abnormal, discrimination in the discriminant analysis, the estimated error rate was 7.5 percent ; in the N/U/C, or normal/ulcer/cancer, discrimination, it was 25 percent. These figures agreed with the results of the cluster analysis and the inspection of the contour by six physicians.
The other approach was extraction of the features through the two-dimensional orthogonal transform of the grey level image of the region of the apex. The selected best set of features for the N/A discrimination resulted in the error rate of 6.4 percent; the best set for the N/U/C discrimination resulted in that of 21.3 percent. These figures also agreed with the results of the cluster analysis and the inspection of the images by an experienced physician.
In both approaches, the selected features tended to belong to the different clusters of the features. It is clearly recognized in the N/A discrimination and supports the validity of the selected features.
The following conclusion may be drawn from the experiments :
1. It has been proved that there exist feature sets of the error rate comparable to that of physicians.
2. While discrimination between normal and abnormal samples was good, that between ulcerated and cancerous samples, on the contrary, was poor.

抄録全体を表示

Visually evoked potentials for stimuli of various monochromatic lights were recorded at three positions in the visual pathway of a monkey brain; namely the optic tract (OT), the lateral geniculate body (LGB), and the visual cortex (VC). Waveforms of these potentials were analyzed applying the principal component analysis, and the opponent color responses were derived at each position. It seems that the 1st and the 2nd scores correspond to the opponent color responses of the Yellow-Blue and the Red-Green components, respectively. These opponent color responses are very similar to those of the human brain.
In the analyzing process of visually evoked potentials, the first and the second characteristic vectors were precisely calculated by using the technique of the singular value decomposition. These characteristic vectors are regarded as carriers of color information, and are supposed to correspond to the basic waveforms which transmit the Yellow-Blue and the Red-Green responses. And the : phase relation between those characteristic vectors is discussed.
The contribution of the 1st characteristic value, which corresponds to the Yellow-Blue response, gradually increases from the OT to the LGB and the VC. This phenomenon might be caused by the enhancement of the blue component, suggested by DeValois et al.

抄録全体を表示

An automated glottal image analysis system capable of processing ultra high-speed laryngeal films successively has been developed.
The objective of this ultra high-speed laryngeal film is to obtain a permanent record of vocal cord motion during phonation. These films are then analyzed frame-by-frame for data relevant to the theory of vocal cord motion and to the detection of laryngeal pathologies. Because of the massive amount of data, several automated systems have been developed to extract such parameters as glottal area, length, width, etc.
The system that has been developed detects the glottal boundary from the glottal image, and describes its shape by angular bend function. And it measures glottal area and width for the glottal waveform. Using the data of glottal shape and glottal waveform, estimations of the vocal cord vibration patterns are attempted. Some parameters are defined ; as “Difference Factor (D)”, “Complexity Factor (F)”, and “Symmetry Factor (S)”. Parameter D is to measure the difference between two glottal boundary shapes. Parameter F is to measure the complexity and width of glottal patterns. And parameter S is to measure the symmetry of vocal cord vibration patterns. By using these parameters, vocal cord vibration patterns can be estimated quantitatively.
In this paper, we analyzed the images of extracted larynx of a dog under several conditions ; such as the control case, additional weight loaded case, silicon injected case, operated case, and the over-operated case. As a result, each condition has been classified clearly, and successful observation of the glottal wave and detection of laryngeal pathologies have been made possible.

抄録全体を表示

The present paper describes an attempt to display 3-D shaded images, which have been derived from a series of RI emission computed tomographic images (ECT) obtained by using a rotating gamma camera or a 7 pinholes collimator.
The creating procedure of the 3-D shaded images is as follows : The contours of the tomographic images are first determined. The 3-D geometric model of the organ of interest, which is approximated by many triangular patches, is summed up to be constructed in a computer. This model is projected on a screen and its hidden surfaces are eliminated by using a priority algorithm. Then the image is shaded to show up the surfaces of the organ vividly.
The system can be implemented in a standard gamma camera data processing system in order to help a physician to recognize the spatial relationship among several tomographic images.

抄録全体を表示

Effects of data compression, to save the processing time in the P-wave recognition program with the adaptive correlating filter (ACF), on the performance of the ACF has been studied.All QRS or QRST regions were eliminated from the record to compress ECG data.
Two indices were adopted as criteria for evaluating the effects. One was derived on the basis of the performance characteristics of the ACF, while the other was the mean error probability in detecting P-waves by using the ACF.
In result, it was shown that the method for data compression could make up for the loss of the P-wave information caused by the data reduction. So the data reduction proposed here makes data compression feasible with little deterioration of the ACF performance.

抄録全体を表示

As a new diagnostic technique for heart diseases, a two-dimensional mapping system of body surface microvibration due to cardiac cycle has been designed and built as a trial. The system is composed of an optical fiber transducer and a microprocessor-equipped display system.
The optical fiber transducer is composed of three parts : (1) He-Ne laser light source, (2) phototransistor sensor and (3) optical fibers which guide the light from source to object and from object to detector. On the tip of the transducer, optical fibers are arranged in a honeycomb lattice shape with proper separations so that the intensity of the reflected light will be proportional to the displacement of the object, and the surface angle to the light axis can be compensated. In addition, the difference of surface reflectivity is also compensated using the light intensity, detected by another optical fiber arranged at a relatively far distance.
The data processing unit of body surface cardiac potential mapping is used for A/D conversion, storage and display of the data.
The preliminary experiments on five healthy male subjects suggested that the normal precordial vibration pattern accompanied a large retraction during systolic period at the right parasternal area.

抄録全体を表示