小橋 昌司

We propose the method for position/pose identification and manipulator navigation to the desired position using one camera and three-dimensional CAD model. In position/pose estimation, instead of simultaneous estimation of all position/pose parameters of target object, the parameters are severally estimated with image features. Thus, our method consists of position parameters estimation, pose parameters estimation and manipulator navigation to the desired position. The position parameters are estimated approximately with image features, and then pose parameters are estimated. And then, a manipulator to the desired position based on the estimates is moved. Both estimation and movement are repeated until the manipulator reaches to the desired position. This gradual approach achieves manipulator navigation concerning automation works with improvement in speed and accuracy.

© 2003 IEEE. This paper defines normality in human body for diagnostic analysis of signs observed in human body. The normality is a matter of degree. Physician usually examines whether a patient is either normal or abnormal. Diagnosis of human body is usually done by observing biosignals, radiological images, body surface information and others of human body. First, the information granularity of these signs of human body is shown. The normality is defined in the theory of hierarchical definability. According to the definition, a calculation method of the degree of normality is introduced. Finally, the examples of the degree of normality are shown.

This paper describes a computer sided support system for localizing the distal transverse screw hole in interlocking intramedullary nailing. It is required to localize the distal transverse screw hole under surgeries for femur fracture. Conventional X-ray CT method can find the hole in the CT table. In it, when we move patients from the CT table to the surgery table, the exact potion is missed. Therefore, a real time determination of the hole position is required to enhance the quality. This paper solves the problem by using an ultrasonic system aided by fuzzy logic. First, we show the details of the equipments. Second fuzzy inferences effectively determine the fuzzy degree being the screw hole inserting to cattle femur. Finally, the experimental result shows that our system can determine the screw hole with 1.0mm of error.

Functional assessments of the meniscus assists to diagnose diseases of the knee joint. To estimate the degree of the degeneration of the injured meniscus, this paper proposes an automated method for segmenting the region of the meniscus from multidetector computed tomography (MDCT) images. The novel feature of the proposed method is that we segment the meniscus based on a fuzzy rule-based representative line finding algorithm. The representative line, which is a new concept, is a line that abstracts the position and shape of the meniscus. Experimental results on the extended and flexed knee joints for three healthy subjects show that our method found a proper representative line, and segmented the meniscus successfully.

This paper presents a rough sets-based method for clustering nominal and numerical data. This clustering result is independent of a sequence of handling object because this method lies its basis on a concept of classification of objects. This method defines knowledge as sets that contain similar or dissimilar objects to every object. A number of knowledge are defined for a data set. Combining similar knowledge yields a new set of knowledge as a clustering result. Cluster validity selects the best result from various sets of combined knowledge. In experiments, this method was applied to nominal databases and numerical databases. The results showed that this method could produce good clustering results for both types of data. Moreover, ambiguity of a boundary of clusters is defined using roughness of the clustering result.

言語優位半球の判定は, これまで後遺症の危険を伴ったWadaテストが用いられてきた.近年, 近赤外分光法(NIRS)の原理を応用した光トボグラフィ装置が, 言語野の解析に適用できる無侵襲な方法として用いられるようになってきた.本文ではファジィモデルを用いた言語優位半球判定法を提案する.本方法は, まず言語野の所属度をファジィモデルを用いて表現し, 言語野の同定作業を行う.次に言語野の所属度に基づいた左右性指標の計算を行う.本文では3例に適用し, 本方法は言語野の同定に有効であることを示した.

functional MRI (fMRI)とは頭部MR画像の連続撮像を行い, MR信号の時間変化から脳機能を解析する技術である.刺激に伴い賦活の生じた領域ではMR信号が上昇し, やがて信号はベースラインに戻る.これをヘモダイナミクスレスポンス(以下HR)という.本文ではウエーブレット変換を用いた賦活領域の同定, HR反応遅延を計測するfMRI解析法を提案する.実験では20例のfMRIデータに本方法を適用した.本方法の賦活領域の同定精度はSPMと同等であり, さらにHRの遅延の計測が可能である.また, 本方法を用いることで補足運動野と運野においてHRの遅延に有意差があることが初めて確認された.

脊椎疾患の手術療法において, 術者が脊椎神経を損傷する危険性があるため, 現在はMRやCTによる脊髄造影画像で術前に脊椎内の神経の状態を観察する.しかしこれらの画像は2次元画像で, 医師が関心領域である硬膜嚢の正確な3次元形状を認識するのは困難である.組織間の形状を術前に把握できる支援システムの開発が望まれている.そこで本研究では, 支援システムの基礎となる腰部MR画像からの硬膜嚢の自動抽出法を提案する.本抽出法は硬膜嚢の形状に関する知識からファジィ論理により硬膜嚢らしさを求める.実験では, 被験者5人(健常者4人, 疾患者1人)から得られたMR画像に対して本方法を適用し, 抽出領域の3次元表示と医師による手抽出の結果との一致度から提案法の有効性を確認した.

© Springer-Verlag Berlin Heidelberg 2002. In medical images, the tube-formed tissues such as the blood vessels, the trachea, and the pancreatic duct are sometimes partially masked because of the constriction, stones in the vessels, the pancreatic cancer, etc. Therefore, it is not easy to automatically segment the region of tubes (ROTs) from medical images for visualizing the structures by using conventional image segmentation methods, because inference of ROTs is difficult. In this paper, we propose a fuzzy rule-based augmented reality method for finding non-continuous ROTs. We can obtain the ROT without extracting it. The physicians’ procedure for finding the ROT can be eliminated by fuzzy inference techniques based on their knowledge. The employed knowledge is the intensity, the curve, and the radius of the ROTs. We apply the proposed method for finding the pancreatic duct from MR Cholangiography images. Through experimental results, we show that this method can successfully find the pancreatic duct from any data sets and it can clearly visualize the 3D shape of the ROT in MIP images.

We propose a new algorithm called representative line detection algorithm embedded physician knowledge with fuzzy if-then rules. The algorithm is to detect the representative line of the region of interest (ROI). The representative line can show the rough location and shape. We first consider the representative line which consisted of some nodes. These nodes are then automatically detected by tracking most suitable direction from the starting node. To evaluate this algorithm, it applies to segmentation of the meniscus from CT knee images. The experimental results of six normal subjects showed that the representative line detection algorithm could express center line of the meniscus, and could lead to detect successful segmentation of the menisci for the all.

Augmented Reality (AR) is a combination of a real scene viewed and a virtual scene generated by a computer. This paper introduces fuzzy inference techniques into AR. The fuzzy inference based AR enhances a region of interest (ROI) in medical images using expert knowledge expressed with fuzzy if-then rules. Therefore, the AR enables users who are not familiar with such medical images to observe the ROIs as well as expert's observation. Especially this paper discusses the fuzzy inference based AR in MR Cholangiopancreatography images. The AR enhances tube-formed tissues such as the pancreatic duct. In this case, a step for finding the tissue using fuzzy inference is included. We apply the proposed the fuzzy inference based AR to three subjects. Through experimental results, we showed the pancreatic duct was successfully found from any subjects and could clearly augment the 3D shape of the pancreatic duct.

This paper describes comparison between our developed two Non-Destructive Testing (NDT) system with ultrasonic method and pulse-radar method. These systems could provide information of embedded objects in concrete structure. First, we introduce these NDT systems. The raw output images of these methods keep us from understanding them. To overcome this problem, we apply fuzzy logic into them. Next, we compare these two developed systems to show how to select appropriate method for various cases. Consequently, pulse-radar system is good for general purpose, because this system needs no coupling medium. Ultrasonic system is appropriate for the purpose (e.g. to detect both embedded objects and cracks, to examine deeper area, and in case of gathering of reinforcing bars). We recommend that these two systems are complementally for realizing NDT system.

This paper proposes an application of a Genetic Algorithms (GA) to Fuzzy Non-Destructive Testing (NDT) System. We are concerned here with the system extracts embedded tubes from pulse-radar images. The system usually uses several fixed parameters to analyze input data. The optimal values of the parameters depend on the environment (e.g. permittivity of concrete, pulse-radar device, and so on). Therefore, the system often fails extract embedded tubes when the fixed parameters are not optimal in the environment. No work deals with this problem. This paper attempts to optimize them by a simple GA in order to solve the problem. We applied the system with the GA to two data sets obtained by various environments. As the result, the system with the GA was able to adapt each environment.

In transcranial sonography system, we usually placed the sensor to anterior and superior of the attachment of the upper ear (posterior temporal window). Due to this limitation, we cannot obtain transcranial information from arbitrary place of the skull. Although this limitation free system is strongly required in clinical treatment, there is no discussion of the placement free transcranial sonography system. This paper describes a new transcranial system without this limitation of the placement. We did the experiment for investigating to identify the tissue under the artificial bone, which assumes the skull and cerebral hemisphere. In it, we employ a fuzzy inference technique to achieve high identification robustness. This system is available for radiation free bedside technique, emergency usage, and non-invasive usage for diagnosing brain diseases.

The information obtained from lacustrine laminated diatomite gives us the environmental records in the past. The varved diatomite is composed of the alternation of light-colored and dark-colored laminae. The lamina thickness has the periodic variation, which is most likely affected by climatic change due to periodic solar activity. In this paper, we propose an effective method for evaluating the space-varying lamina thickness of lacustrine diatomite. The proposed method is achieved by using 2-dimensional(2D) filters and space/spatial-frequency analysis. Through the simulation of lacustrine laminated diatomite, we show that this simple imaging algorithm allows clear and intuitive display of the change in the lamina thickness with respect to space (the distance from the surface).

Near-infrared spectroscopy (NIRS) is a recently developed method, which can investigate the human brain function with noninvasive, high time resolution, and high portability. It has been applied to assessment of language dominance by Watanabe et al, however, it has a risk which some measurement points took other brain activities occurred in like the motor area. This paper improves the method by introducing fuzzy logic. Fuzzy logic is used to giving fuzzy degrees belonging to the speech, motor and other areas for each measurement point. Using the obtained fuzzy degrees, we redefine the laterality index as the fuzzy mean of the laterality indexes for all paired measurement points.

In this paper, we describe a development of a new ultrasonic oscillosensor and its application to recognition of human action. The ultrasonic wave is used in many fields due to its low cost and non-invasiveness. The ultrasonic wave sensitivity catches the difference of the sound property and it propagates in the water. By using this, we develop a new oscillosensor using the ultrasonic wave. Firstly, this paper shows the principle of this new sensor. Secondly, as an example of the practical application, this paper describes a recognition system of human action on the medical bed. From the result, this sensor can monitor the vital action of a patient and it can detect a state that a patient falls into bad condition, such as motionless or suffering.

Assessment of language dominance is one of important pre-treatments of brain surgery. The objective of this paper is to propose an estimation method of 'Laterality Index (LI)', which is a measure of language dominance, using near-infrared spectroscopy and imaging (NIRS/I) modality. The proposed method called fuzzy LI is a fuzzy-extension of the conventional estimation method (Watanabe et. al., Neuroscience Letters, 1998) called crisp LI. The benefit of our method is that the fuzzy degree belonging to language area for each measurement point is given, and that the language area is dynamically determined study by study using only NIRS/I. Experimental results on seven normal healthy right-handed volunteers showed that the fuzzy LI was correlated with one's handedness investigated by Edinburgh score at 77%.

This paper shows an automated method for segmenting the lateral ventricles from human brain MR images. It enables us to automatically volumetry and construct 3-D rendering images. First of all we segment the whole brain and the cerebrospinal fluid (CSF) by using a 3-D MR image processing software developed by Hata et at Our method for segmenting the lateral ventricles is based on fuzzy inference techniques, which is able to represent expert's knowledge and to introduce the knowledge to image processing The employed knowledge on the lateral ventricles is the location, the intensity, and the shape The proposed method was applied to MR volumes of 20 normal volunteers, 20 Alzheimer disease and 20 hydrocephalus patients The experimental results validated that this method was able to segment the lateral ventricles with high accuracy.

Volume visualization of cerebral blood vessels is highly significant for diagnosis of the cerebral diseases. It is because the automated segmentation of the blood vessels from an MR angiography (MRA) image is a knotty problem that there are few works on it. This paper proposes an automated method to segment the blood vessels from 3D time of flight (TOF) MRA volume data. The method consists of: (1) removing the background, (2) volume quantization by watershed segmentation, and (3) classification of primitives by using an artificial neural network (NN). In the proposed method, the NN classifies each primitive, which is a clump of voxels, by evaluating the intensity and the 3D shape. The method was applied to seven MRA data sets. The evaluation was done by comparing with the manual classification results. The average classification accuracy was 80.8%. The method also showed the volume visualizations using target maximum intensity projection (target MIP) and surface shaded display (SSD). The evaluation by a physician showed that unclear regions on the conventional image were clearly depicted on applying the method, and that the produced images were quite interesting for diagnosis of cerebral diseases such as aneurysm and encephaloma. The quantitative and qualitative evaluations showed that the method was appropriate for blood vessel segmentation.

Functional magnetic resonance imaging (fMRI) provides high-resolution datasets that allow neuro-scientists to sensitively detect activated region relating to a given task. Cerebral hemodynamic response (HR) of neural activity at the activated region is delayed and dispersed in time. This study proposes a new data-driven analysis for detecting the activated region and measuring the temporal delay of HRs. Our method can be divided into four steps. Firstly, a standard preprocessing (realigning and high pass filtering) is applied to the fMRI data. Secondly, we reconstruct the time course of signal in order to improve the time resolution. Thirdly, we derive the form and the delays of HRs using continuous wavelet transform. Finally, we give wavelet coefficient as the degree of activation. The proposed method has been applied to 7 right-handed subjects on right or left hand-gripping task. Experimental results showed that primary motor cortex (M1) and the supplementary motor area (SMA) were detected as activated region, and means of temporal delays of HRs were right/left = 4.49/4.92 sec at SMA and 5.41/5.46 sec at M1. Consequently, our method could classify the difference of HR delay between M1 and SMA.

Surface rendering of the endorrhachis is one of the effective techniques to diagnose diseases of the endorrhachis caused by spine malformations. It requires a procedure for segmenting the 3D endorrhachis from many MR images consisting the whole human body, however, few automated extraction methods have been developed before. In this paper, we propose a fuzzy rule-based approach for extracting the endorrhachis from MR lumbar images. Our target portion, the endorrhachis, consists of the spinal cord and the nerve roots. This method is based on knowledge of location and knowledge of intensity. They are expressed by fuzzy if-then rules and compiled to a total degree as the measure of extraction. Our extraction method can be divided into two steps. Firstly, the method extracts the spinal cord by using thresholding and labeling. Secondly, our method extracts the nerve roots with the knowledge derived from anatomical information. The experimental results showed that our method successfully extracted the target portion, and that the results were useful to diagnose the compression of the endorrhachis.

This paper presents a computer-aided diagnosis (CAD) system, which be able to segment the whole brain, the brain portions, the cerebrospinal fluid (CSF) and the lateral ventricles from human brain MR images, and then to give the volumes and the 3-D volume renderings. In the system, the whole brain and the brain portions are automatically segmented by using previously developed software (Hata et al, IEEE SMC-C, 2000). In this paper, we introduce a new concept of a representative line to segment the CSF and the lateral ventricles. The representative line is automatically detected by inferring its direction with evaluating the location in the head, the position in the CSF, and the shape of the line. The inference is preformed with fuzzy inference technique. The developed system was applied to MR volumes of 20 normal subjects, 20 Alzheimer disease and 20 hydrocephalus patients. The segmentation error ratio was 1.98% in comparison with the volumes of manually delineated region. This allows us to experimentally characterize modes of variation that are indicative of disease processes.

Near-infrared spectroscopy (NIRS) is a recently developed method, which can investigate the human brain function with noninvasive, high time resolution, and high portability. However, there are few discussions on post-processing of time series data taken by the NIRS because of the difficulty of understanding the obtained data and the complexity of the human higher-order brain function. This paper discusses on an analysis of such a time series. The analysis method is based on fuzzy c-means (FCM) clustering and wavelet transform, and it divides the time series of a measurement point into some clusters with respect to wavelet coefficients. To evaluate the performance of the method, it has been applied to four healthy volunteers, and three brain-dead patients. The results showed that the proposed method could segment the NIRS time series into some clusters that may represent brain states, and could estimate the number of clusters.

This paper proposes a computer aided diagnosis system of meniscal tears from 3D human knee MR image with T1-weighted and T2-weighted MR images. The first step of our method is a 3D image registration between both images on the computer display by manual. The second step determines the candidate region of the menisci from T2-weighted MR image aided by fuzzy if then rules with respect to the location and the intensity. The final step determines the voxels in the menisci from the candidate region by using T1-weighted and T2-weighted MR images. We applied this method to several subjects. All voxels in the menisci of each subject were successfully identified and their 3D surfaces were displayed. Thus, our developed system would improve to diagnose the meniscal tears.

<p>現在の医療現場において、MRIやCT装置の利用は必要不可欠な診断手段である。それに伴い計算機による診断支援システムの構築が行われてきたが、画像診断は専門家である医師の知識に負う所が多いため、容積測定や立体表示の自動化は困難であった。そこで、本研究では医師の知識をファジィインフォメーショングラニュレーション(ファジィIG)の概念に基づいてファジィモデル化した医用画像処理法を示した。本論文は、第2章:頭部MR画像からの全脳領域抽出の為のファジィIGによる自動しきい値発見法、第3章および第4章:ニューラルネットワークおよびファジィIGを用いた頭部MR Angiography画像からの脳血管領域抽出法から構成される。さらに第5章で提案した医用画像処理法を核とした脳機能診断支援システムを構築し、実際の臨床例を用いてその有用性を検証した。本研究では、ファジィIGを医用画像処理に応用することで、個人差や症例差が激しい臨床例にも適用できた。</p>

Hemispheric dominance for language has been assessed by the Wada test which is an invasive technique with considerable risk of complications. Recently, optical topography system has been used as a non-invasive method which is applicable in detecting the language dominance. Our method can be divided into three steps. Firstly, preprocessing is applied to the data. Secondly, we estimate the degree of the speech area. Finally, we calculate laterality index based on the degree of the speech area. Our results demonstrate that our method is effective for decision of language dominance.

This paper proposes an automated discriminating method of material using ultrasonic imaging that can conform change of tissue elasticity. The proposed method employs fuzzy infernce. Firstly, three attributes are acquired from ultrasonic wave of the known material. Secondly, membership functions are constructed using three attributes automatically. Finally, we presume the material using the function from unknown ultrasonic wave. We apply this method to two situations with bone phantom and without bone phantom. The experimental results on six materials showed that each presumed rates under the two situations were 86.6% and 63.3%.

This paper proposes an automated procedure for segmenting an MR image of a human brain based on fuzzy logic, An MR volumetric image composed of many slice images consists of several parts: gray matter, white matter, cerebrospinal fluid, and others. Generally, the histogram shapes of MR volumetric images are different from person to person. Fuzzy information granulation of the histograms can lead to a series of histogram peaks. The intensity thresholds for segmenting the whole brain of a subject are automatically determined by finding the peaks of the intensity histogram obtained from the MR images. After these thresholds are evaluated by a procedure called region growing, the whole brain can be identified. A segmentation experiment was done on 50 human brain MR volumes. A statistical analysis showed that the automated segmented volumes were similar to the volumes manually segmented by a physician. Next, we describe a procedure for decomposing the obtained whole brain into the left and right cerebral hemispheres, the cerebellum and the brain stem, Fuzzy if-then rules can represent information on the anatomical locations, segmentation boundaries as well as intensities. Evaluation of the inferred result using the region growing method can then lead to the decomposition of the whole brain. We applied this method to 44 MR volumes, The decomposed portions were statistically compared with those manually decomposed by a physician, Consequently,our method can identify the whole brain, the left cerebral hemisphere, the right cerebral hemisphere, the cerebellum and the brain stem with high accuracy and therefore can provide the three dimensional shapes of these regions.

This paper shows an application of fuzzy information granulation (fuzzy IC) to medical image segmentation. Fuzzy IG is to derive fuzzy granules from information. In the case of medical image segmentation, information and fuzzy granules correspond to an image taken from a medical scanner, and anatomical parts, namely region of interests (ROIs), respectively. The proposed method to granulate information is composed of volume quantization and fuzzy merging. Volume quantisation is to gather similar neighboring voxels. The generated quanta are selectively merged according to degrees for pre-defined fuzzy models that represent anatomical knowledge of medical images. The proposed method was applied to blood vessel extraction from three-dimensional time-of-Right (TOF) magnetic resonance angiography (MRA) images of the brain. The volume data studied in this work is composed of about 100 contiguous and volumetric MM images. According to the fuzzy IG concept, information correspond to the volume data, fuzzy granules corresponds to the blood vessels and fat. The qualitative evaluation by a physician was done for two- and three-dimensional images generated from the obtained blood vessels. The evaluation shows that the method can segment MRA volume data, and that fuzzy IG is applicable to, and suitable for medical image segmentation.

This paper presents a fuzzy rule-based region growing method for segmenting two-dimensional (2-D) and three-dimensional (3-D) magnetic resonance (MR) images. The method is an extension of the conventional region growing method. The proposed method evaluates the growing criteria by using fuzzy inference techniques. The use of the fuzzy if then rules is appropriate for describing the knowledge of the legions on the MR images. To evaluate the performance of the proposed method, it was applied to artificially generated images. In comparison with the conventional method, the proposed method shows high robustness for noisy images. The method then applied for segmenting the dynamic MR images of the liver. The dynamic MR imaging has been used for diagnosis of hepatocellular carcinoma (HCC), portal hypertension, and so on. Segmenting the liver, portal vein (PV), and inferior vena cava (IVC) can give useful description for the diagnosis, and is a basis work of a pre-surgery planning system and a virtual endoscope. To apply the proposed method, fuzzy if-then rules are derived from the time-density curve of ROIs. In the experimental results, the 2-D reconstructed and 3-D rendered images of the segmented liver, PV, and IVC are shown. The evaluation by a physician shows that the generated images are comparable to the hepatic anatomy, and they would be useful to understanding, diagnosis, and pre-surgery planning.

This paper proposes an architecture of a registration system for medical images. Image registration is the process of determining correspondence between all points in two images of the same scene, and is now widely used to medical images. In medical imaging, segmentation, registration and interpolation play primary roles. In those, registration is the most time consuming task because we must compare all voxel data and then evaluate the matching degree many times. Quantitative evaluation criterion of matching degree with multiple-valued coding of the image feature is proposed, and all architecture to save the processing time of the data comparison is described Finally, as a practical application, we described the summary of a registration of human brain MR volume data to diagnose brain disease.

The paper introduces registration systems of multi-modality medical images and describes the practical systems related to brain science. A possibility for applying soft computing techniques is also shown. First we describe a registration system of computed tomography image and magnetic resonance angiography image of a human brain. This registration system is used to demonstrate anatomical location information of vascular lesion from the surface of the human skull. We next describe a registration system of magnetic resonance (MR) image and positron emission transmission (PET) image. The MR image can produce neuroanatomical information, and the PET image quantifies metabolic pathways in vivo. In both systems, we describe a possibility of soft computing techniques.

This paper shows a novel medical image segmentation method applied to blood vessel segmentation from magnetic resonance angiography volume data. The principle idea of the method is fuzzy information granulation concept. The method consists of 2 parts: (1) quantization and feature extraction, (2) iterative fuzzy synthesis. In the first part, volume quantization is performed with watershed segmentation technique. Each quantum is represented by three features, vascularity, narrowness and histogram consistency. Using these features, we estimate the fuzzy degrees of each quantum for knowledge models about MRA volume data. In the second part, the method increases the fuzzy degrees by selectively synthesizing neighboring quantums. As a result, we obtain some synthesized quantums. We regard them as fuzzy granules and classify them into blood vessel or fat by evaluating the fuzzy degrees. In the experimental result, three dimensional images are generated using target maximum intensity projection (MIP) and surface shaded display. The comparison with conventional MIP images shows that the unclarity region in conventional images are clearly depict in our images. The qualitative evaluation done by a physician shows that our method can extract blood vessel region and that the results are useful to diagnose the cerebral diseases.

This article shows a SD-reconstruction method of dynamic MRA images of the liver. The method is based on fuzzy if-then rules that represent expert's knowledge. The membership functions used are determined using accumulative histogram of given images. Thus, each of steps runs automatically with robustness. In the experimental result, the method is applied to three dynamic MRA data sets of the liver. Qualitative evaluation by physicians were done for reconstructed 3D images. It shows that the method segments the hepatic veins precisely, and given images are essentially for surgery planning.

This paper describes useful fuzzy logic techniques for medical image segmentation. Specific methods to be reviewed include fuzzy information granulation, fuzzy inference and fuzzy cluster identification. Fuzzy information granulation is introduced as a powerful scheme to find the thresholds to obtain the whole brain region in MR data. Fuzzy inference technique succeeds to segment the brain region into the left cerebral hemisphere, right cerebral hemisphere, cerebellum and brain stem. The fuzzy inference aided segmentation procedure is also useful to human foot CT image. Fuzzy cluster identification is adapted to determine the obtained clusters into blood vessel or other tissues in MRA image.

PURPOSE: To determine the features of cortical atrophy in frontotemporal dementia (FTD) and Alzheimer disease by using a hemispheric surface display generated with magnetic resonance (MR) images. MATERIALS AND METHODS: The extent of cortical atrophy was evaluated with automated MR hemispheric surface display and volumetry in 18 patients with FTD and in 18 matched patients with Alzheimer disease. Results were compared with those in 18 healthy, matched control subjects. RESULTS: Most cortical regions, were significantly atrophic in FTD and Alzheimer disease. The frontal and anterior temporal lobes were significantly more atrophic in FTD than in Alzheimer disease. The mean hemispheric-to-intracranial volume ratio in patients with FTD (56.2%) and those with Alzheimer disease (58.4%) was significantly smaller than the ratio in the control subjects (66.0%). Asymmetry of hemispheric volume was significantly larger in the FTD group than in Alzheimer disease and control groups. CONCLUSION: Cortical atrophy in FTD is more widespread than was previously thought. Asymmetric frontal and anterior temporal atrophy is a distinctive feature of FTD and distinguishes it from Alzheimer disease. Hemispheric surface display is a useful complement to tomographic images and is useful for the evaluation of focal cortical atrophy in degenerative dementia, especially FTD.

本文では, ファジィインフォメーショングラニュレーション(fuzzy Information Granulation; fuzzy IG)による, 頭部MR画像からの脳領域抽出の為の濃度しきい値発見法を提案する.Fuzzy IGとは, Zadehが提唱する新しい概念である.ファジィ情報(fuzzy information)は, 幾つかのファジィグラニュール(fuzzy granule)で構成され, fuzzy IGによってfuzzy informationからfuzzy granuleが引き出される.本手法では, fuzzy informationを濃度ヒストグラムとして, fuzzy granuleを脳を構成する白質部, 灰白質部, 脳脊髄液のヒストグラム上のピークとする.Fuzzy IGは, 医師の知識をモデル化したヒストグラムと, 画像の濃度ヒストグラムとのファジィマッチングによって行う.得られたしきい値を用いて自動抽出した脳体積と, 手動で得られた体積とを, 50例のMR画像に対して比較した結果, 平均誤差率は2.3%であった.

This paper shows an application of fuzzy information granulation to medical image segmentation for MR angiography image. The method treats medical image as information, and the granules are the anatomical parts such as blood vessel, fat, etc. Our granulation method consists of two stages : (1) Quantum generation and (2) Integration of them to produce granules. Quantum generation is performed with watershed segmentation. Integration process is achievedby estimating the degree for predefined models. By granulating MR angiography image, we obtain the bloodvessel region. In the experimental results, we show the target M IP and shaded surface display images of the segmented region. They are useful for evaluating the lesions such as the areury.

One of the most serious problems in medical science is an increment of Alzheimer&#039;s disease. It is known that the patient&#039;s brain atrophy is a result of neural cell loss. It is useful for the diagnosis of Alzheimer&#039;s disease to measure the volumes of the brain portions and to display them. We can obtain anatomical information from 2D slice images produced by MRI. We propose a computer-aided system for the diagnosis of Alzheimer&#039;s disease. The system consists of: (1) extraction of the portions from MRI data, (2) measurement of the volumes of the portions and then displaying them, and (3) user interface for a medical doctor. In this paper we describe procedures for the above. For extraction of the brain portions we propose the method based on standard region growing algorithm and the method of figure decomposition using the distance value. Comparison of the volumes of our extracted portions with volumes manually measured by a physician shows that the error rate, on the average, is 1.74% for 48 MRI data. We also discuss the 3D display, the measuring range, and the construction of the user interface for a physician. © 1997 Scripta Technica, Inc.

MR脳画像同士の位置あわせは, 同じ脳疾患を持つ複数の患者の脳画像を重ねあわせることによって, 濃度値などの情報から共通する部位の異常を視覚的に捉えることが目的である. しかし, MR脳画像は大局的に見ると一様に見える部分であっても, 局所的には個々により異なる部分が多く, このことがMR脳画像同士の位置あわせを困難としている. 従来, 画像同士の位置あわせの方法がいくつか紹介されている. それらは, 画像の濃度値による相関関係による方法, 輪郭の曲率や変極点の位置情報による方法などである. 相関関係による方法のような対象画像の濃度値をもとに行われる方法は, 濃度値情報を取得するというMR脳画像を位置あわせする目的から使用することができない. また, これらの方法は対象画像の局所的な一致度に基づいて全体の位置あわせを行うものであり, 局所的に多くの部分が異なる脳画像の位置あわせに適用するのは困難である. 本文では, ラバー (ゴム) をモデル化した物理モデルを提案し, このラバーモデルを用いてMR脳画像の位置あわせを行う. ラバーモデルとは, 平面状のゴムを想定し, このゴムに歪みが加わったときのゴム全体の力学的な振る舞いをモデル化したものである. ラバーモデルでは局所的な変位が画像全体に影響を及ぼす構造をしており, 全体として協調しながら画像全体が移動するため, 大局的な位置あわせが可能である.

近年, MRやCT装置などの医療機械の発展によって, 人体内部が簡単に撮影できるようになった. また, 最近のコンピュータの高速化, メモリの大容量化によって, それらの情報を解析・可視化し, 評価することが容易になってきた. 一般的には, 興味ある部位をその医療画像から抽出し, それを評価・可視化・解析することが必要である. その際, 抽出目的部位の濃度幅が異なるため単純なしきい値処理では抽出できないのが現状である. そこで, 本稿では, 複数存在する抽出目的部位毎にしきい値処理を行って得られた画像群を, クリーネ代数を用いて統合する方法を示す. 更に, クリーネ代数を用いても解が得られない場合に対して, Brzozowskiの演算を用いることで解が得られることを示す.