小橋 昌司

This paper describes a dentistry support system for root canal treatment using ultrasonic. Presently, no complete method of root canal treatment has proposed. Dentist empirically removes the dental pulp using measures of length of root canal. Therefore, the support system for root canal treatment is required to precisely remove the dental pulp. This paper solves the problem in root canal treatment by our ultrasonic device and fuzzy logic techniques. We determine the dentin-dental pulp junction by calculating two fuzzy degrees of amplitude of the echo and approximate distance of the surface and the junction. As the result, our system can determine the thickness of dental pulp within error of 0.289 mm.

This paper proposes a fuzzy rule-based system for estimating the surface roughness by ultrasonic waveform. Our estimation method consists of three steps. The first step extracts characteristic values from an object with known surface roughness. The second step constructs fuzzy membership functions with respect to characteristic values. At the final step, the fuzzy rule-based system with estimates the surface roughness of an object with unknown surface roughness. Moreover, we propose a method for removing noise of the result by using characteristic value of ultrasonic echo. We applied this method to phantom with three kinds of surface roughness. Then, our system can successfully estimate the roughness of the phantom.

In this paper, we propose a real-time position and pose tracking method for pin insertion to a pipe-like hole in almost uniform motion region. In the conventional methods for estimating position or pose of the region, a three-dimensional (3D) model data of the target region is often needed. For reducing the processing time, our approach achieves the position and pose estimation without the model data. In order to estimate and track it using a single camera, we set some landmarks on the region surface. Additionally, Kalman filters are used to estimate 3D position of the landmarks. We apply the proposed method to the insertion system for a pipe-like hole in a solid object. In the experimental results using a manipulation robot, we show that it can insert a stick into the hole in the almost uniform motion region in real-time.

The vibration of human vital activity has usually 10Hz or less. The sensor that effectively detects this vibration is required in the field of medicine and nursing. In this paper, we propose an extraction system of sleep state using a new ultrasonic oscillosensor. This sensor has 30dB S/N in the resonance frequency and the size of 26mm diameter x 10mm height, and it has a linear characteristic in a band being lower than the resonance frequency. The resonance frequency of this sensor is 6.8Hz for longitudinal and transverse vibrations. Thus, this sensor effectively detects low-frequency vibrations. We propose a system that detects the sleep state by using this sensor. In the experimental results, we extracted the number of flop-over at the 13% of mean error ratio and extracted the sleeping time at 3.5% of mean error ratio for the healthy volunteers. For the two patients: one is insomniac and the other is not so, we were able to extract the difference between the patients.

Applications as visual navigation of mobile robot with image sensor and mixed/augmented reality have been investigated actively. Many of these techniques require the localization of human or robot. In this paper, we propose a novel method for estimating the position of a camera by using edge in time-series of images. The technique can be applied to indoor environment. In our approach, the camera position in next frame is predicted from the estimated trajectory of the moving camera in known environment. Both the correction based on the prediction and current measurement, and matching frameworks with the edge of background images, result to efficient estimation of 3-dimentional camera position.

In this paper, we propose a real-time method for analyzing 3-dimensional(3D) scene of objects from time series of monocular images. A single camera observes some points on the object so that 3D position of the points can be estimated by extended Kalman filter. We apply this method to two real-time applications. One is to acquire 3D geometry of an object, and another is to estimate 3D pose/position of an object. This approach needs no model data of the object a priori and achieves the estimation of 3D geometry and pose/position. In the experiments using a manipulation robot, we show that it is effective method to estimate 3D information with high accuracy.

This paper describes a design method of automated medical diagnosis system (ANMS), which provides a normal degree for a disease. In this paper, we consider blood test data of human. Suppose a disease which can be diagnosed by the test, first, we do statistical analysis. Second, we determine the reference range of the test. Third, we design a fuzzy inference system consisting of membership functions based on the reference range. The inference system plays primary role in the AMDS. Finally, we show the design of AMDS for diabetes and the experimental results.

X-ray fluoroscopic images are widely used for evaluating the knee kinematics after the total knee Arthroplasty, TKA in short. There are some conventional methods for estimating the motion of the knee implant in vivo using single-plane fluoroscopic images and 3-D geometric models of the knee implant. However, these methods have a problem that estimation results are dependent on the initial pose and position given at the start frame of sequence for analysis because these methods analyze the motion of a series of the knee implant for every frame based on the pose and position estimated at the previous frame. This paper proposes a 3-D/4-D registration method that estimates the initial pose and position by predicting dynamic knee kinematics. To quantitatively evaluate our method, it was applied to computer-simulated images and phantom images that took the knee implant in vitro fixed with arbitrary pose and position by a jig. The experimental results showed that the proposed method could estimate the pose of within the error of 0.04 deg in the computer-simulated images and 0.82 deg in the phantom images, and the accuracy was improved in comparison with the conventional methods. © 2005 IEEE.

A form of a support implant used in revision total hip arthroplasty (THA) should be diagnosed periodically because it may be distorted or broken. This paper proposes a new method that estimates the strength of stress on the support implant using finite element method (FEM) and genetic algorithm (GA). 3D sectional images of the support implant in vivo are acquired from multidtector-row computed tomography (MDCT) devices. The proposed method searches for a model whose shape is same as the support implant in MDCT images using GA. The database consists of many parts of the support implant that are generated by FEM with various patterns of strain. Direction and strength of the stress are estimated from the searched model. As a result of the proposed method applying to simulated MDCT images of horseshoe-shaped phantoms, the proposed method found the correct model. Also, the method detected the strain within the error of 1.03±0.38mm and 33.0±12.3N. Additionally, we applied the method to MDCT images of 2 patients after THA. The results showed that the direction and the strength of the stress were successfully estimated. © 2005 IEEE.

This paper describes combination rule of normal degrees in human body in automated medical diagnosis system. The normal degree is defined in a framework of fuzzy logic. Physician usually examines whether a patient; is either normal or abnormal for a disease. The, normal degree is calculated in automated medical diagnosis system. The practical examples of medical images and blood test are described. In it, it is shown that union or inter-section operators are introduced for calculating normal degrees on MR, meniscal tear images and blood test for diabetes.

人の肺は葉間裂と呼ばれる膜状の組織によって区切られた，５つの肺葉により構成されている．CT画像を対象として肺を肺葉単位に分割することは肺の機能解析や，病変の診断等に対して有益な情報を提供することができる．我々の以前の研究では，管状組織密度と抽出した葉間裂に基づいて手動で入力した初期平面から最適な肺葉境界を推定する手法を提案した．この手法によって誤差1.5mm程度の精度で肺葉境界が推定できた．この誤差の多くは肺葉境界の付近において発生している．本研究では以前に提案した肺葉分割のシステムに対し，ノード位置を固定していたために推定精度が低かった肺野境界上に存在するノードを定義し，それらの位置を更新する事，メッシュの構造を変更する事，またノード間の滑らかさの表現を変更することにより，肺葉分割のアルゴリズムの改良を行う．

視覚機能を持ったロボットや, 複合・強調現実感に関する研究が盛んである.これらの技術の多くは, 人間やロボットの自己位置姿勢情報が必要である.本報告では時系列画像中のエッジ情報を用いて, カメラ位置を推定する方法を提案する.本方法は屋内の環境に適用され, 既知環境中でのカメラの軌道情報に基づき次フレームでのカメラ位置姿勢を予測する.予測した位置情報に基づき, 環境フレームモデルをエッジ抽出した入力画像へ投影し, その投影モデルを修正しながら, 抽出エッジとのマッチングを行い, カメラの3次元位置を推定する.

人工股関節全置換術(THA ; total hip arthroplasty)後における再置換術により人体に埋め込まれたサポートインプラントにおいて,術後の経年劣化や日常生活での荷重により生じた形状の歪み・折損は,日常動作の妨げや痛みを与える危険性がある.そこで本論文では,Multidetector-row computed tomography (MDCT)を用いて生体内のサポートインプラントの形状変化を3次元可視化する計算機診断支援(CAD ; Computer-aided diagnosis)システムを提案する.本システムではMDCT画像からサポートインプラント領域をファジィ画像処理により抽出し,3次元形状データとの形状認識・歪み検出の2段階マッチングを行うことで,表面同士の距離値に基づき歪み度合いを算出し,これを3次元可視化する.

実環境を撮影した画像から撮影時とは異なる視点の画像(自由視点画像)を生成する研究に関し,複数台のカメラを用いて実現する方法が多く提案されている.カメラを複数台用いる方法に代えて,一台のみによる自由視点画像生成が可能になれば様々な分野で応用可能と考えられる.そこで本報告では,一台のパンチルトカメラを用い,そのカメラパラメータと撮影される背景の情報を用いて,その環境内にある未知物体の3次元形状情報や位置情報を推定,またその自由視点画像を生成する方法を提案する.

実環境を撮影した画像から撮影時とは異なる視点の画像(自由視点画像)を生成する研究に関し,複数台のカメラを用いて実現する方法が多く提案されている.カメラを複数台用いる方法に代えて,一台のみによる自由視点画像生成が可能になれば様々な分野で応用可能と考えられる.そこで本報告では,一台のパンチルトカメラを用い,そのカメラパラメータと撮影される背景の情報を用いて,その環境内にある未知物体の3次元形状情報や位置情報を推定,またその自由視点画像を生成する方法を提案する.

実環境を撮影した画像から撮影時とは異なる視点の画像(自由視点画像)を生成する研究に関し,複数台のカメラを用いて実現する方法が多く提案されている.カメラを複数台用いる方法に代えて,一台のみによる自由視点画像生成が可能になれば様々な分野で応用可能と考えられる.そこで本報告では,一台のパンチルトカメラを用い,そのカメラパラメータと撮影される背景の情報を用いて,その環境内にある未知物体の3次元形状情報や位置情報を推定,またその自由視点画像を生成する方法を提案する.

人工膝関節置換手術後,生体内の人工膝関節の動作機能評価にはX線透視画像が広く用いられている.従来様々な,X線透視画像を用いた動作機能評価法が提案されてきたが,Occlusionが発生している画像への適用は困難であった.そこで本研究では,Occlusionが発生している画像に適用可能な人工膝関節の動作機能評価法を提案する.提案手法は,X線透視画像と人工膝関節の3次元形状データを用い,ファジィイメージマッチングにより,人工膝関節の位置・姿勢を認識し,動作機能を定量化する.本提案手法をシミュレーション画像,ファントム画像,被験者の画像に適用した結果,Occlusionが発生した画像においても,動作機能評価に有効であることを確認した.

Unconstrained health monitoring systems have received much considerable attention in medical applications, because such system can examine a subject without constraint. In this study, we propose a detection method based on a fuzzy logic for evaluating heart rate using our ultrasonic vibrograph. In the experiment for confirming heart rate, our method has been successfully used to detect the heart rates of four subjects, compared with a method using all electrocardiograph.

痴呆性疾患は大脳全体が同じ割合で萎縮するのではなく,痴呆ごとに萎縮する脳葉に違いがあることがわかっている.そのため,大脳を区分し,前頭葉,頭頂葉,側頭葉,後頭葉単位の体積を測定することは痴呆症の鑑別,経過観察に非常に有用である.本稿では,各葉間の境界をfuzzy rule-based ASMを用いて最適な形状に変形し,3次元頭部MR画像から各葉を自動的に区分する方法を提案する.再現性の評価のため,提案法を3人の被験者に適用しユーザ毎の体積のばらつきを調べた.1人のエキスパートと3人の初心者は各被験者につき1つの3次元画像を10回ずつ区分した.その結果,提案システムを用いることによって高い再現性で大脳を区分できた.

これまでFuzzy C-means (FCM)、Bayesian Classificationなどの様々なクラスタリング法が医用画像の領域分割に用いられてきた。しかし、ほとんどのクラスタリング法は数値データのみを対象とし、名義データを扱えないもしくは扱うのに数値化が必要であるため、領域分割への応用においても、数値で表現できる画素値などの情報しか使えなかった。そこで、本研究では名義データを効率的に扱えるラフ集合論による画像の領域分割法を提案する。本提案法では、しきい値処理とラベリングにより求めたラベル番号をクラスタリングの属性値とする。すなわち本提案法では、従来の画素値に基づくクラスタリングではなく、画素の連結関係に基づくクラスタリングを可能とする。また本文では本手法の性能を評価するため、提案手法を頭部MRI画像に適用し、その真値画像と比較し数値的な評価を行う。

Investigating the shape and time to peak (hemodynamic response [HR] delay) of the HR introduces a new facet of research on human brain function. This paper proposes a novel method for analyzing event-related functional magnetic resonance images (ER-fMRI). The method is based on wavelet transform. The principal feature of the method is that it can detect all activation area and estimate the HRF simultaneously. To evaluate the proposed method, experiments were done on phantom data and human subjects. The results showed that the method could detect the activation area equivalent to and more than SPM99. Then, examination of variability of HR delay among the activation areas suggests that HR delay at the motor area is significantly faster than that at the visual area. (C) 2004 Published by Elsevier B.V.

This paper describes an ultrasonography system of implant in holies aided by fuzzy logic. Although conventional X-ray CT method can visualize the implant in the CT table, it has serious disadvantage of X-ray exposure. Therefore, we require the system which call be used the orthopedic Surgery with no X-ray exposure. This paper solves the problem in an intramedullary nail by our ultrasonic device and Fuzzy logic techniques. We localize the distal transverse screw hole in the nail under Surgeries for femur fracture. Their screw hole positions are identified by calculating two fuzzy degrees of intensity and the variance. As the result, our experimental result Shows that Our system call identify their screw hole positions within error of 2.5 mm.

This paper introduces rough sets theory into image segmentation problems. The proposed image segmentation method considers that an is an information system and a pixel in the image is a datum of the information system. The feature values of the pixels reflect connectedness between pixels. With respect to connetedness. the proposed method extracts alpha-tolerance relation, which is an extended concept of rough sets. from the information system. The categories of the alpha-tolerance relation mean regions in the image. Consequently for each region pixels have high connectedness to other pixels of the region. To evaluate the performance of the proposed method, it was applied to MR images produced by the BrainWeb computer simulation. The experimental results showed that the proposed method segmented the given MR image with high homogeneity.

We propose a method for estimating the pose/position of a guiding reference object using both camera images and 3-dimensional model data of the guiding object. Our objective is to identify the object position/pose and to move an optical pointing device (laser pointer) mounted on an active camera so that LED light is pointed at the desired orientation and position. During a surgical procedure, probe positioning should be achieved at precisely desired position and orientation. For this purpose, we update input images of a camera on the pointer and iterate estimations by moving the camera. This iterative approach with updated images makes the pose/position parameters to be estimated more accurately than approaches with only one image. Though iterated estimation generally increases the computational effort, estimation task is achieved with improvement in speed by the approach with changing image resolution. Through some experimental results, we show that the proposed method is effective for guiding system.

Forced vital capacity (FVC) of individual lung lobes is important information for thoracic surgeries such as the living-donor lobar lung transplantation. However conventional methods for measuring the FVC of individual lung lobes have not considered the variation of the lobar function among subjects. This paper proposes a new method for measuring the FVC of individual lung lobes using, Multidefector-row CT. The method first estimates contribution ratios of individual lung lobes for the FVC of whole lung. Using the estimated contribution ratios and the FVC of whole lung measured by a spirometry, our method estimates the FVC of individual lung lobes. As a result of applying the proposed method to live normal subjects, we could successfully measure the FVC of individual lung lobes.

We propose a novel method for detection of a moving object in all image sequence. This purpose is achieved by using discrete wavelet transform and extended complex Kalman filter(FCKF) in the transform/spatio-tcruporal mixed-domain. The moving objects are considered as arbitrary trajectory signals(ATS). A trajectory signal in mixed-domain is a bunch of 1-dimentional complex signals. And all ATS is a bunch of quasi-periodic signals with pitch fluctuations in mixed-domain. A parallel bank of ECKFs is employed for the estimation of quasi-periodic signals. In addition, these estimates call be combined with depending of the energy Of file subband images. Experimental results show the moving object is detected effectively.

In this paper, we propose the novel method for 3D shape reconstruction of object using an extended Kalman filter with a single active camera. In conventional methods, the laser scanner or stereo camera is often used as the sensor to reconstruct 3D shape. However, they have large-scale system and some problems caused by it. We use only one active camera for shape reconstructing. Since an active camera takes timeseries images, some points can be observed and 3D position of the points estimated by extended Kalman filtering. Also, we consider the reconstruction of 3D geometry from the connection between each point, and plan good camera's viewpoints. After estimating 3D position of two selected points on the object and analyzing them, the active camera moves to the next viewpoint for obtaining the hidden information. By using these estimates of points from some planned viewpoints, 3D shape reconstruction is achieved. We apply the proposed approach to the images of computer simulation and real world, and we show that it is effective for shape reconstructing of object.

This paper describes an automated medical diagnosis system (AMDS), which determines a normal degree for a medical sign or disease based on fuzzy logic. First, we summarize normal degree on theory of hierarchical definability (THD) by Zadeh. Second, we show flow chart of AMDS, which consists of two parts; one is construction of AMDS and the other is calculation of normal degrees. Third, we demonstrate two examples of AMDS: one is AMDS for diagnosing Alzheimer disease and Frontotemporal dementia and the other is personal AMDS for diabetes. Finally we conclude technical results.

This paper proposes a real-time position and pose tracking method of a moving object. The conventional methods need a three-dimensional (3D) model data of the target object. By the proposed method, the 3D position and pose tracking can be achieved without the 3D model data. We set some landmarks to estimate the 3D position and pose of the object by a single camera. In addition, the trajectory of moving landmarks is estimated by a Kalman filter. The experimental results show that a stick with a diameter of 5.0 mm can be inserted in a moving pipe with a diameter of 30.0 mm in real-time.

In conventional transcranial sonography system for human, available window of skull in adults is limited to the temporal bone. Visualization of an intracranial tissue front all angles can support to diagnose the brain disease with high accuracy. However, no transcranial sonography system has been developed. This paper describes transcranial ultrasonography System based on anatomical knowledge. Generally, it is difficult to visualize intracranial tissues from frontal bone because of large attenuation in skull. To overcome this difficulty, we employ fuzzy signal processing and human brain anatomical model. In our system, we can successfully extract the skull and the brain surface for human. As a result, our method could visualize two-dimensional skull and brain surface.

In this paper, we propose the novel method for estimating the three dimensional position and pose of the object by using only one active camera. Our method estimates the depth length from the camera to the object by a Kalman filter. Kalman filter enables the depth distance estimation with high accuracy by using one camera and decreases the number of the iteration by estimating the accurate depth distance. Moreover, we iterate the estimation by updated input images. Although the distance from a camera to the object was about 600 mm, the experimental results showed that the estimated error of the depth length was about 2.4mm.

In this paper, we propose a method to generate arbitrary viewpoint images of unknown objects using a single camera set in a known environment. in recent years, various methods for generating arbitrary viewpoint images have been proposed. These methods often use multiple cameras. When we can generate the arbitrary viewpoint images using only a single camera, we can apply the technique to more various fields. Therefore, we propose the arbitrary viewpoint image generation method by using a single camera and the information of the known environment (size of a room, and the relation between the room and the camera). We set the camera at the rectangular parallelepipeds room and unknown objects on the floor. We generate the arbitrary viewpoint images using the estimated three-dimensional (31)) information of the unknown objects and the known environment information. In the experimental results, we show that the proposed method can estimate the 3D information of the unknown objects and generate the arbitrary viewpoint images.

Investigating the shape and time to peak, called HR delay, of hemodynamic response (HR) gives a novel tool to analyze a human brain function. We have proposed an analysis method for detecting activated areas and measuring HR delay simultaneously. In this paper, we improve the performance of the method by modifying a mother wavelet. The modified mother wavelet is based on mathematically defined HR function by SPM (statistic parametric mapping) software by Friston et al. To evaluate the proposed method, the proposed method and conventional methods are applied to phantom data and human subjects. The results indicated that the proposed method could detect activated areas with equivalent to and more accurately than the conventional methods.

Investigating the shape and time to peak (HR delay) of hemodynamic response (HR) introduces a new facet of researches on human brain function. We have proposed an analysis method for detecting activated areas and measuring HR delay simultaneously using event related function magnetic resonance images (ER-fMRI). The method has been based on wavelet transform. In this paper, we investigate the performance of the method by modifying a mother wavelet. The modified mother wavelet is based on mathematically defined HR function by SPM (statistic parametric mapping) software by Friston et al. To evaluate the proposed method, we applied the proposed method and conventional methods to the phantom data. The results showed that the method could detect the activated area equivalent to and more than conventional methods. We also estimated the accuracy of measuring HR delay. The results showed that the proposed method could measure HR delay more accurate than the conventional method.

The human lung is composed of five compartments called "Lung lobes." Segmenting lung lobes from chest Multidetector-row Computed Tomography (MDCT) images can provide effective information for analyzing pulmonary conformation and detecting pulmonary diseases such as the emphysema. Our method finds the boundary surface between lung lobes by the fuzzy control that fits a curved surface into the space with lower density of tubular tissues, which are defined as peripheral blood vessels and bronchus. This paper describes the performance evaluation of the proposed method and the application method for estimating the forced vital capacity of individual lung lobes.

This paper proposes an automated 3D visualization method of embedded tubes applicable to the scanned result of pulse-radar Non-Destructive Testing (NDT). The proposed method consists of three stages. First, our method defines the processing region which includes a pattern generated by a tube. This region is determined by referring to the composition of a received wave. Second, after expert knowledge of a tube is translated into fuzzy inference rules, the positions of embedded tubes are identified by inferring them. Third, 3D links of the identified positions are formed to visualize the continuous shape of the tubes. Consequently, the tubes are extracted, and their 3D shapes are visualized. The experimental result on the specimens shows that our method was able to find all tubes that exist in the radiograph and the schematic. Our method could thus provide the internal information of concrete with sufficient accuracy required in the practical construction work.

This study introduces a novel method for enhancing MR cholangiography (MR-C) images based on representative line finding algorithm. The method enhances the images using the representative line (RL) of the pancreatic duct. The RL is tracked by finding the most suitable direction that is decided by means of fuzzy inference techniques. The tracking performance was evaluated by computer simulation, and by applying our method to MR-C images of tube-formed phantoms. The results showed that the proposed method could correctly find the RL of the objects. The diameters of the tube-formed phantoms were estimated with high accuracy. Then, the proposed method was applied to enhance MR-C images of four normal volunteers and one patient. In all cases, the pancreatic duct was enhanced clearly.

We describe a new automated method for detecting embedded objects in the ultrasonic non-destructive testing (NDT) system. A-scan waves collected by our developed system are converted into a B-scan image. The sensor system has the noise signals independent from targets to be detected. In the ultrasonic NDT system, the signals are due to disturbing of echoes produced by the transducers and multiple reflections. These signals are called inherent wave. This paper first proposes the estimation method of the inherent wave from the B-scan image. After this method subtracts the inherent wave, the resultant image (suppression image) is considered as the image consisting of only echoes from the embedded objects. Second, analysis of the intensity histogram of the suppression image leads the candidate points of embedded objects. Finally, fuzzy if-then rules can represent information on distribution of the intensity histogram and the homogeneous intensity levels of the objects. Evaluated degrees from the inference results can demonstrate the embedded objects. The method was applied to concrete members with reinforcing bars, resin tubes and steel pipes. The experimental results showed that this method was able to automatically detect the embedded objects with high accuracy and to display the location of embedded objects.

This paper proposes an automated method for finding the Willis ring from the human brain MR angiography (MRA) images, which can depict cerebral arteries with high contrast. It strongly helps screening of unruptured cerebral aneurysm in MRA images. The proposed method consists of (1) segmenting cerebral arteries from MRA images, (2) skeletonization of artery trees, and detection of furcations, and (3) finding furcations in the Willis ring using genetic algorithm (GA) based on fuzzy knowledge base (fuzzy KB). Fuzzy KB gives knowledge about the Willis ring that consists of arteries and furcations. GA finds a set of furcations by optimizing an objective function. The objective function used by GA estimates fitness of a set of furcations using fuzzy KB. Our method was first applied to a 3-D phantom data generated by computer simulation. The result demonstrated that our method detected all suitable furcations correctly. Next, it was applied to MRA volume data of two normal healthy volunteers. In any cases, the proposed method detected desired all furcations in the Willis ring.

Periodic inspection for degradation of insulating oil in transformers requires for preventing outage. However, conventional methods were not suitable for that, due to a large size of the measurement system and requirement of sampling. This study aims to develop the ultrasonic evaluation system satisfying these requirements aided by fuzzy inference. The system evaluates degradation, and it estimates kinetic viscosity. In the experiment we applied our method into six lands of insulating oils. As a result, con-elation coefficient between our estimation and viscometer were over 0.92. Thus, our method correctly determines degradation of insulating oil.

Periodic inspection for degradation of insulating oil in transformers requires for preventing outage. However, conventional methods were not suitable for that, due to a large size of the measurement system and requirement of sampling. This study aims to develop the ultrasonic evaluation system satisfying these requirements aided by fuzzy inference. The system evaluates degradation, and it estimates kinetic viscosity. In the experiment, we applied our method into six kinds of insulating oils. As a result, correlation coefficient between our estimation and viscometer were over 0.92. Thus, our method correctly determines degradation of insulating oil.