Super-Kamiokande is a 50,000ton cylindrical water Cherenkov detector located 1000m underground. The detector was viewed by 11,146 photo-multiplier tubes with 50 cm diameter which cover the 40% of the inner surface. The primary aim of this experiment is to measure cosmic neutrinos and to search for proton decay. In 1998, by the detailed study of the atmospheric neutrinos, the neutrino oscillation was discovered and it was proven that the neutrinos have finite masses. For the future large water Cherenkov detector, the development of the new light sensor is necessary.

View full abstract

According to the increase of requirements for the surface quality of steel sheet for car outer panels, a new automatic on-line inspection system for fine defects, called “pattern-flaw”, has been developed. The system can certainly detect harmful pattern-flaws by distinguishing them from harmless patterns, oil and processing liquid remainder, by adopting the principle of the 3-channel polarizing camera detection of polarized ligiht reflection from the steel sheet. It is confirmed that it achieves 100% of detection rate and less than 1% of over-detection rate as a result of the on-line comparative examination with the human inspection.

View full abstract

This paper describes a 3-D shape measurement system for objects with various reflectance properties. Although triangulation-based 3-D shape measurement methods have been widely used for many industrial applications, generally only Lambertian objects can be measured with the methods. In order to apply this type of measurement method to specular objects, it is necessary for the image sensor to detect the incident position and the angle of the light stripe. However, it is difficult to detect the incident angle of the light stripe using conventional image sensors. Thus, we here devised a sensor which has the capability of detecting the incident angle directly, and then proposed a new 3-D shape measurement method using this sensor. The experimental results demonstrate that the system can measure the shapes of hybrid objects with almost the same level of accuracy as it can those of Lambertian objects and specular objects.

View full abstract

Conventional Brillouin based strain sensors, whilst advantages over FBG-based sensors for structural monitoring applications, have low spatial resolution and are inadequate for dynamic strain measurements due to their large measurement lengths of several minutes. We present a correlation-based continuous-wave technique for high spatial resolution and distributed dynamic strain measurements. Using our technique, we have successfully measured dynamic strain from a 5cm vibrating section, at a sampling rate of 18.4Hz, and having strain accuracy of about 45.2με.

View full abstract

The rate of plastics being recycled in our country has become increasingly high. Especially the rate for PET (polyethylene telephtalate) bottle recycling is showing rapid increase. To further increase the recycling rate of plastics, not only the progress of the reproduction technology of plastics but the development of an identification technology will be indispensable. To date, we have reported a plastic identification technology using laser diodes. In order to put this technology into practice, we studied how the forms of the plastics will affect the identification results. As a result, it turns out that by using our technology it is possible to discern the composition of plastics without being influenced by their form. In this paper we report especially on identification of PET bottle.

View full abstract

In the previous report we proposed the spectral matching imager, which, using the correlation image sensor, detects an object having a particular spectral function out of a scene under a typical frame rate. In order to increase power efficiency of the illumination, this paper proposes a type of spectral matching imager that uses AM-Coded multispectral illumination instead of the variable wavelength monochrome illumination used previously. The AM-coded multispectral illumination projects a white light each wavelength of which is amplitude-modulated with a different frequency. Using a reference signal created as a weighted average of the AM waveforms of the illumination in which the weights are the values of a reference spectral function sampled at corresponding wavelengths, the imager outputs pixelwise spectral correlation between the object and the reference, due to orthogonality of the sinusoidal AM waveforms. The theory and an implementation of the imager are described.

View full abstract

Recent trend of micro mechanical sensors based on silicon micromachining is described. Pressure, force, acceleration, angular rate and other mechanical quantities can be measured using the micro sensors. Wafer process packaging plays an important roles for low cost, reliable and small size sensors. High performance rotational gyroscope based on electrostatically levitated micro motor, electrostatically levitated spherical three-axis accelerometer, small size fiber optic pressure sensor and extremely sensitive resonation force sensors are described.

View full abstract

With development of silicon micro-machining technology, R&D and commercialization of micro mechanical sensors progressed. International standardization is indispensable to production, sale, and international trade. The organization and the deliberation method of IEC international standardization were described. As for a semiconductor pressure sensor, an international standard was published in 2001, and a semiconductor accelerometer is under deliberation internationally now. In addition to these, the micromachine terminology was proposed to IEC recently.

View full abstract

Recent progresses of the Micro Total Analysis Systems (μTAS) appeared in previous μTAS symposium are reviewed. Main interests are sifted from DNA to proteomes and biological cells in biochemistry applications. High performance and high sensitivity sensing methods which are indispensable in μTAS have been currently studied. Microfluidics in the microchannel and micro flow device have been also developed focusing actual applications. Micro flow cell for bio-sensing applications and biomolecles handling systems are also described.

View full abstract

Human body shape data have become more and more utilized for designing well-fitting products. The following specifications are required for the measurement equipment of human body to obtain data easily used for the designing purpose: (1) the accuracy should be 1.0 mm for the whole body, 0.5 mm for the head; (2) measurement duration should be within 1.0 sec in order to minimize the body sway effect; (3) hidden areas should be minimal. We have developed new scanners based on these specifications. The head scanner is developed by NEC Corp., and the whole body scanner by Hamano Engineering Co., Ltd. The head shape including back of the ear and under the chin could be measured by the head scanner within 1.0 sec. The body shape including the crotch, the armpit and under the bust could be measured by the whole body scanner in 1.8 sec.

View full abstract

Make-up simulation based on face recognition technology can be considered as a useful application for users to find charming facial expression. In order to describe vivid facial expression, movie-based make-up simulator that performs like a mirror is required because the facial expression is always changing. The make-up simulator allows to put lipstick, eye shadow, eye liner, powder on the face, and rouge on the cheeks. Moreover eyebrow make-up and hair color changes are also contained in the simulator. These make-up image processing based on facial parts tracking runs on normal PC and DV camera in real-time. Consequently the real-time make-up simulator enables the users to try various types of cosmetic products without actual make-up.

View full abstract

We have developed a robust and accurate method to detect facial feature points to realize a real-time and interactive digital make-up system. In addition to the feature point detection method, important features, such as object contours and a face direction, are also detected. Especially face direction is effective information for cheek area that cannot be described by definite contour obtained from 2-dimensional images. We have integrated the above-mentioned methods to a simple digital make-up system.

View full abstract

Against the background of remarkable progress of computer technologies, large-scale systems have been complicated. While, the human interfaces have been diversified, for instance a CRT (Cathode Ray Tubes) device and a touch panel operating system. On account of the diversification, many human errors occurred caused by the consistency between human recognition and interfaces. In order to execute effective improvement of this situation, there is a demand to establish a human interface evaluation method. The authors have developed two methods to estimate human error rate and evaluate human interface quantitatively. In this paper, they explain their outlines and applications.

View full abstract

In this paper, we propose a three-dimensional position detection method for moving object using multiple HyperOmni Visions (HOVIs) and a new surveillance system for old person living alone based on the proposed method. The system consists of some HOVIs and an active camera. The system detects the position of a person using HOVIs. When the person stops, the active camera shoots the person’s image and the system recognizes the person’s conditions. Experimental results showed the feasibility of this system.

View full abstract

Lip Reading is known as one of the word recognition method in high noise environments. We propose a hardware realization of lip reading system by using FPGA. By hardware realization of a lip reading system, high-speed and compact stand-alone realization is able to be expected. In this paper, configured for real time processing is expected and operation of each part in the system is conferred. Experimental results of lip reading are also included and are compared with ordinal system developed as software program to show the effectiveness of the proposed system.

View full abstract

This paper presents a display system that enables us to find necessary information quickly. Multiple windows containing information appear from the bottom of the display, moving upward swiftly. We move our eyes to obtain the information in the windows. The speed of the eye movements controls the upward moving speed of the windows so that we can get information from the system as fast as possible. If we find anything interesting in a window and stop the eye movements to look at it more carefully, the system stops the window motion, enlarging the window to show details.

View full abstract

We proposed a method of 3D caricature generation which is based on the automatic extraction of the facial parts from 3D facial image. This method is likely to suffer sometimes fatal degradations in the feature extraction from a variation of the head pose (roll, pitch, yaw). Therefore, we propose a method of facial image modification in rotation (roll) which is based on the irises position extracted by Hough transform of the circle from gray scale image. We improved the facial feature extraction rate and the quality of the caricature generated by the improved 3D face data.

View full abstract

We studied the method for objectively evaluating emotion through FST (Facial Skin Temperature), brings less stress and non-contact measurement. According to Russell, emotion can be complained by pleasant level and arousal level. Then, we tried to construct Arousal level estimate model using FST. We applied local fractal dimension analysis to the FST. We examined relation between the temperature distribution of a nose and α wave amplitude of EEG, reflect arousal level. As result, the local fractal dimension analysis of nasal temperature can be adapted for evaluating temporary arousal level.

View full abstract

The purpose of this paper is to construct the individual identification system by the facial skin thermogram (FST). FST gives a constant image under any illumination condition. Then, we extract the individual feature from FST with the image processing and examine the evaluation method to identify individual. We adopted neural networks with back-propagation (BP) learning algorithm. The identification experiment was done among 60 (20×3) registered patterns and 250 (50×5) input patterns, consequently the correct acceptance rate was 73% and the correct rejection rate was 73.3%.

View full abstract

We propose a novel system for real-time three-dimensional surface orientation measurement. The advantages of our method are: 1) single frame capture of normal vector distribution, 2) dense, pixel-wise capture of normal vectors, and 3) independence of background illumination and surface reflectance nonuniformity of the object. This system consists of two components; one is sinusoidally amplitude-modulated three-phase (3P) light sources at vertices of a triangle and the other is the 3P correlation image sensor (3PCIS) for demodulating the amplitude and phase of reflected light on the surface. Based on the photometric stereo principle, the phase and amplitude can be easily converted to the orientation and inclination, respectively, of the surface. We implemented this system using our CMOS 32×32 pixel 3PCIS and successfully reconstructed the normal vector map. We also show several experiments for applying this system to personal identification by face.

View full abstract

We have proposed a new classifier which is able to handle multi-class discrimination problem very easily. The classifier consists of Boosting (AdaBoost) units and Fuzzy Partition(FP) unit. The architectre of Boosting + FPM leads to not only promotion of classification precision but also real-time classification without degrading precision. We have applied Boosting + FPM to classify face images to four class (puerility, youth, middle age, advanced age) As a result, Boosting + FPM gives more than 79 and FPM without Boosting gives 63% and 48% respectively. In addition to that, Boosting + FPM gives real-time classification which precision is not less than 72%.

View full abstract

In this paper, we propose a classification-based approach for locating human faces in cluttered images. We extract gradient features from local image and the dimensionality of the feature vector consisting of gradient features and gray scales is reduced by the PCA for the sake of computation efficiency and detection efficacy. The polynomial neural network (PNN) takes as inputs the binomials of the projection of the feature vector onto the subspace learned by PCA and is trained on face and non-face samples to discriminate the two classes of patterns. In order to further improve the detection performance, we incorporate into PNN the distance of feature vector from the subspace (DFFS). The experimental results on testing a large number of images with complex backgrounds demonstrated the efficiency of the proposed method in terms of detection performance and computational cost.

View full abstract

Recently, human verification methods which use physical characteristics (biometrics) command public attention. The face verification is the one method. The usual face verification methods used the 2-D information such as light and shade image and contour image as a main stream. But it couldn’t get satisfactory verification rate if much constraint wasn’t put on users in the case of the data acquisition. And because they were poor at change in illumination surroundings and the correction of the face direction couldn’t be done. Therefore, we proposed a method which measures the 3-D face shape by light-section method and corrects direction of the 3-D face shape information. Then it verifies by using feature quantity that normal vector on the surface of the face is expressed as a complex number. In this paper, detail of the proposed method and result of the verification experiment to discuss effectiveness of the method are shown.

View full abstract

Gyroscopic Force Measuring System (simply called GFMS) is an entirely new force sensor for measuring a force vector in 3-dimentional space. This paper reports about the characteristics of our prototype GFMS made on first trial. It is able to measure the only one component of a force vector, excluding servomechanisms to estimate angles of incidence. The feedback gains are selected somewhat arbitrarily, but the fundamental experiments show that she force 0.3 [N] can be measured by the GFMS. This suggests the possibility to measure a small force range useful for airflow distribution in an air-conditioned room as an example of expected applications. No doubt the experimental results are to be released in a future publications.

View full abstract

The hysteresis characteristic, which is caused by body material, body form, strain gauge, adhesives, etc., is included in one of the output characteristic of a load cell. And it has become the main factors of a measurement error. We create the function that approximates a hysteresis error characteristic curve to compensate a hysteresis error correctly by combining two or more functions. We hit on an idea of the method of carrying out error compensation and approximated this characteristic correctly. We enforced this method to the Roberval mechanism load cell and acquired the desirable result for attaining the required accuracy of measurement equipment.

View full abstract

As uncertainty contributions for the static torque calibration of a torque measuring device, the following components are considered; repeatability with unchanged mounting position, reproducibility with changed mounting position, zero error, interpolation error, reversibility (hysteresis) and resolution including the ability of the indicator. Some foreign standard documentations, however, have not revealed any experimental confirmation for the validity of assumed probability distribution for each contribution. In order to establish a determination method of uncertainty for the torque calibration, the validities of probability distributions were verified by the numerous repetitions of torque measurement using the primary torque standard machine. The determination method is proposed based on the experimental results.

View full abstract

A changes with passage of time of calibration weights affect uncertainty of measurement on mass standards and balances. Specially, interval of calibration for mass standards needs to estimate its change with passage of time. We have verified the change with passage of time for our mass standards, 1 mg-200g, three times in the past; February 1996, April 1999 and November 2001. So, this paper introduces its detailed analysis with measuring uncertainty.

View full abstract

Force standard machines (FSMs) of Japan consist of deadweight type, hydraulic amplification type and lever amplification type (LM). The LM has an advantage of compactness, however, it has a drawback of long term instability of the amplitude ratio due to abrasion of knife edges caused by repeated calibrations. Therefore, periodic comparison with other FSMs is necessary to assure the performance of the LM. We made a comparison of the LM with other FSMs and verified the performance of the LM.

View full abstract

This work is concerned with the development of a signal processing algorithm for checkweighers to realize much higher speed of operation and highly accurate measurement of mass of object during crossing a conveyor belt. In this paper an employed digital filter is of Finite-duration Impulse Response (FIR) type that can be designed under the consider ation on the dynamics of checkweighers. The experimental results on checkweighers suggest that the filtering algorithm proposed in this paper is effective enough to practical applications.

View full abstract

Fingers of human allow an object to be lifted using adequate grasping force and without slippage, even though the weight and friction coefficient of the object are unknown. Grasping force is controlled by detecting complex response of the tactile receptors. In this study, we develop a distributed sensor to lift an object by fingers without producing excessive force. Geometry of sensor is designed by calculating the contact condition between the finger and a plate with tangential load using a FE (finite element) analysis.

View full abstract

This paper describes characteristics of fluxgate sensors with troidal permalloy (Fe-Ni: alloy) core. It has been revealed that excitation conditions such as frequency and current for troidal primary coil and the output voltage provided by the detection coil are remarkably affected by the number of core layers and primary coil inductance. We have ascertained experimentally the optimal number of core layers and primary coil inductance. A high sensitivity against external magnetic field has been verified connecting the detection coil of the optimized sensor core and coil assembly to the signal processing system of the fluxgate magnetometer.

View full abstract

Paired comparison method is one of the most effective methods to scale an ambiguous quantity in the sensory evaluation. It is relatively easier to compare and choose one from two samples than to choose directly. Good results can be obtained easily even when examinees are not specialists. However, the number of comparison becomes an issue of this method. It is required to compare n(n-1)/2 times to make the complete paired comparison table. When the number of the samples becomes larger, the number of comparison becomes too huge to compare all pairs. A band function model paired comparison (BMPC) method is one of the paired comparison methods. Since BMPC method uses inequalities to solve the paired comparison table, not all comparison is needed. In this paper, the new sequential solution method to solve paired comparison method is proposed. By the method, not n(n-1)/2 times compares are required to solve n samples paired comparison. The validity of proposed method are confirmed by simulation, and it becomes clear that it is possible to set scale values to each samples without 30% comparison on avarage. We applied the proposed method to the actual pitch comparison experiment. Consequently, we succeeded to obtain the result which is absolutly same as complete result by an incomplete paired comparison by using the sequential comparison method.

View full abstract

In this paper, we tried discrimination of five herb species and mixed gases using SOM (Self-Organizing Map). Previously we used BP method (Back Propagation method) three times hierarchically to discriminate five herb species (Linden, Chamomile, Rose hips, Lavender, Lemon bamu). This time, the improvement of discrimination method was done by using SOM, which is simple method, compared with the BP method. The obtained discrimination result is 96.7 [%]. The discrimination of methane-hydrogenmixed gases was also tried by SOM. This discrimination result is also good. Moreover, as the formed map is reflecting the influence between methane and hydroge gas, and we can see the discrimination result visually.

View full abstract

A periodic pulse train from an ultra fast mode-locked laser can be described as a frequency “comb” of spectral lines in the frequency domain, equally separated by the pulse repetition rate. The optical frequency comb can be used in optical frequency measurements as an optical frequency “ruler”. We have measured optical frequencies of an iodine-stabilized Nd: YAG laser at 532 nm and an iodine-stabilized He-Ne laser at 633 nm. We also applied this technique to an optical frequency measurement of an acetylene-stabilized laser in the telecommunication region using a two-color mode-locked fiber laser.

View full abstract

With the recent development of optical communications system, optical wave-meters with a higher accuracy of frequency are required to test these systems. Therefore, optical frequency standard is ideal for calibrating these optical wave-meters. We have developed frequency stabilized light source using HCN absorption lines. The light source may be used as an optical frequency standard. This paper describes outline of the light source and results of frequency measured by a frequency measurement system developed at National Metrology Institute of Japan (NMIJ/AIST).

View full abstract

We introduce an optical frequency comb (OFC) generator based on an EO modulator installed in an optical cavity. The OFC generator has been employed for measurements of laser frequencies. For that purpose, the span of OFC was expanded to reach to 3OTHz around 1.5 μm by using self-phase modulation in an optical fiber. We propose here a method to generate a high power and flat spectrum of OFC. By this method a multi-wavelength light source for DWDM system can be generated from OFC generator.

View full abstract

Precise evaluation of pulse timing jitters in ultra-short optical pulses is an issue of practical importance for such applications as high bit rate optical communication, optical frequency metrology, optical sampling measurement and so on. The author developed time domain techniques for measuring pulse timing jitter of optical pulses with repetition frequency from 10 Hz to 80 GHz. These techniques have enabled the evaluation of timing jitter over 9 decades of Fourier frequency (2.5 mHz - 18 MHz) with 300 dB dynamic range. This paper describes the principle of techniques and the results of timing jitter measurement for a mode-locked solid-state laser and mode-locked semiconductor laser diodes.

View full abstract

Melanoma is one of the many sorts of tumors which appear on the skin, and it is the most malignant one. Early detection and treatment are very important for curing of melanoma. But there are no objective methods to measure its malignancy and ingravescence. At present the diagnosis of melanoma is mainly done based on the experience of each doctor, and they need some objective measures for diagnosis of melanoma. So we propose the methods to measure its malignancy and ingravescence from the shape and the color information of the lesions with image processing.

View full abstract

Sarcoma is a kind of tumor. There are many types of sarcomas. It is very difficult for physicians to classify types of benign tumor or malignant tumor precisely, except for physicians who specialize in sarcomas. For example, dermatofibroma (benign tumor) and dermatofibrosarcoma protuberans (malignant tumor) are non-epithelial skin tumors which need to be distinguished. In this research, first of all, segmentation of cell nuclear regions whose features are general points of pathologic diagnosis is considered. Using some features computed from the segmented nucleus regions, an automated classification of benign tumor or malignant tumor is finally operated.

View full abstract

The fundus image is used for diagnosing the condition of fundus of the eye. Since observed through the pupil, it is distorted by the optical system of the eye. For an effective diagnosis, reconstructing 3Dfundus surface from multiple fundus images and diplaying is useful. But, we need individual parameters of eye optics and relative camera poses for the reconstruction. For this problem, here, we propose to employ the factorization method as these preliminary steps, and report its feasibility.

View full abstract

Coded aperture CT is one of emission CT techniques. It uses a gamma camera equipped with a coded aperture in place of a ordinary parallel collimator, and reconstructs three dimensional distribution of radionuclides from a projected data collected only from one direction. We have improved the reconstruction image by an algorithm using non-regular-intervals sampling which takes into consideration the difference between the reconstruction range on calculation and the view of an actual sensor.

View full abstract

In order to realize a wearable power assisting suit for assisting a nurse caring a patient in her arm, the power supply and control systems have been miniaturized. The new suit consists of shoulders, arms, waist and legs units and is fitted on her body. The arms, waist and legs have pneumatic rotary actuators driven directly with micro air pumps. The muscle forces are sensed with a new muscle hardness sensor utilizing a sensing tip mounted on a force sensing resistor. This paper gives the design and characteristics of the new power assisting suit verifying its practicability.

View full abstract

The author presented an air mattress method te measure the human bio-information by a non-restrictive manner easy and kind to patients. If we can find some quantitative relations between the bio-information acquired by the air mattress sensor and the sleep stages judged from the polygraph data, the quality of sleep can be estimated easily and kindly. We built a simple mathematical model that provides the estimation of sleep stages only from the heart rate and body motion data acquired from the air mattress sensor.

View full abstract

The authors previously developed an unconstrained and noninvasive measurement system of respiration and heartbeat for drivers using a wire-type respiration pick-up. However, the measurement accuracy of instantaneous heartbeat period was sometimes deteriorated due to the second and third harmonic components of the respiration signal. Adding to it, the measured instantaneous respiration period became sometimes longer than a true value because of a low frequency bias caused by a body twist of the driver at an intersection. This paper proposes a more stable and accurate measurement system for heartbeat and respiration by taking into account the two harmonic components and the low frequency bias.

View full abstract

We evaluated the correlation between CVRR (coefficient of variance of R-R interval) and RSA (respiratory sinus arrhythmia) which show the driver’s heart rate variability, and their subjective fatigue during driving. As the result, we found a correlation between CVRR during driving and driver’s subjective fatigue, and between RSA during driving and their subjective fatigue. Therefore, the fatigue can he quantitatively given as the terms of heart rate variability. These results suggest that we can apply the indexes CVRR and RSA to the security system for the prevention of a traffic accident.

View full abstract

The recent developments of automobile performance and the increase of number of automobiles have lead the national industrial economy. But it also causes the serious automobile accidents called sometimes traffic war. The main cause of the traffic accident, which is one of the social problems, is human-error by drivers. In this paper, due to the fact that more than 90% information necessary to driver a car, we at first carried out the analysis of the basic visual characteristics as from a stand point of the cognitive science. As the result, we found in the visual functions the possibility that a driver causes a traffic accident.

View full abstract

In recent years, because of the deteriorating public order, importance of home security system has been increasing year by year. Considering these points, home security system of novel type is proposed in this paper. The proposed system is constructed with one personal computer as the system server and cellular phones as the clients, and is able to be easily constructed by the user privately. To protect home from intruders, the system automatically takes pictures of intruders and sent them to user by E-mail.

View full abstract

A technique for biometric identification by using hand vein patterns has been demonstrated. The equipment uses a near infrared light source and a monochrome CCD to produce a contrast enhanced image of the subcutaneous veins. We use Phase-Only-Correlation method to ensure the person himself. Studies on several adults indicate promise for security certification uses.

View full abstract

Keeping the life in secure state is one of the important items in the modern society. Unfortunately, modern people must pay effort to keep their life secure. The conventional security system includes variety of security sensors for variety of dangers. This fact leads to the system expensive and complicated. In this research work, we presented one sensor security system. We employed a small condenser microphone as the sensor and investigated how widely it can be applied for variety of dangers. The symptoms by variety of dangers are easily discriminated by applying filters because the changes in the pressure due to the symptoms have the vibrations in the different frequency ranges.

View full abstract