ABSTRACT(english) in Medical Imaging Technology

Abstracts of Medical Imaging Technology of 1999.

January(Vol.17, No.1)

[-> return]

title	author	page
Optical Coherence Tomographic Imaging and Clinical Diagnosis	Naohiro TANNO·Shoji KISHI	3
Key words : Coherence, Interferometer, Reflectometry, Optical coherence tomography, Ophthalmology, Retina, Macula
Abstract : We present a novel method of optical tomographic imaging and clinical diagnosis that are successfully developed in intense light scattering medium of a living organism, using optical time domain short-coherence inteference technology. Construction of an interfermeter with a low coherence light source and a beam scanning yields two-dimentional reflection tomography with high spatial resolution of up to 10μm. A practical case powerfully displayed in clinical diagnosis of ophthalmology is presented. Optical coherence tomography(OCT) is a new diagnostic technology which allows a cross-sectional image of the retina, non-invasively in living eyes with fine resolution comparable to histologic section of light microscopy. Although the fundus can be examined at semi-histologic level a biomicroscope, histopathologic study has been hard to apply it because of rare chance of retinal biopsy or enucleation. OCT revealed the histopathology of macular edema, holes, vitreoretinal interface disease and macular dystrophies in vivo. OCT proved to be a powerfull tool to evaluate the surgical outcome of macular disorders at histological level. Because OCT image is constituted from tissue reflectivity of infrared beam, precaution is required to interpret the cross-sectional images.

[-> return]

title	author	page
Three-dimensional Microscopy Using Femtosecond Pulse Lasers	Satoshi KAWATA, Osamu NAKAMURA	11
Key words : Two photon absorption, Laser scanning microscope, 4 Pi confocal microscope, Pulse laser, Near-infrared, Three-dimensional imaging, Femtosecond
Abstract : We report the principle, fundamental, and applications of two-photon laser-scan fluorescence microscopy using femtosecond pulse lasers. We also present the state-of-the-art technologies with the two-photon microscopy, such as 4 Pi confocal superresolution microscopy, realtime observation of the dynamics of the biological samples using micro-lens array and a regenerative amplifier, long depth-of-focus imaging, and realization of diffraction-limited resolution in ultraviolet region using visible lasers. Future of the related technologies are also discussed.

[-> return]

title	author	page
Biological Imaging Using X-Ray Phase Information	Atsushi MOMOSE	18
Key words : Phase contrast, X-ray interferometer, Cancer, Tomography, Mammography
Abstract : Research activity on phase-contrast imaging in the hard X-ray energy region has been growing recently. Several kinds of methods have been reported using an X-ray interferometer, Fresnes diffraction, and a crystal angular analyzer. The meanings of contrast obtained with those methods are different one another. One can find a merit, however, that as a rule sensitivity achieved by phase-contrast X-ray imaging exceeds that of the conventional X-ray imaging depending on X-ray absorption contrast. Therefore, biological tissues can be observed without the need for staining. Furthermore, high sensitivity is desirable for biological imaging because X-ray dose can be reduced. In the first half of this report, characteristics of those methods are summarized. In the second half, concentrating on the interferometric method which we study, possible imaging targets are discussed for applications to medical diagnosis. Furthermore, our approach to the medical applications from instrumental aspects is also described.

[-> return]

title	author	page
Ultrasonic Imaging by Computing	Yasutaka TAMURA	26
Key words : Computerized ultrasonic imaging, Synthetic aperture, Three-dimensional ultrasonic imaging, Beam forming, Array signal processing
Abstract : We describe the ability and problems of ∪Ultrasonic imaging by computing∩. Computerized indirect beam forming enables us to overcome unavoidable restrictions in conventional direct beam forming techniques. The major implementation problem is a bottleneck in the interface between array signal and digital data. Some computer based beam forming systems including a 3-D ultrasonic imaging are introduced to demonstrate the possibility of a breakthrough in the problem.

[-> return]

title	author	page
In vivo ESR Spatiotemporal Measurements to Analyze the Dynamics of Free Radicals in Living Systems	Tateaki OGATA, Hiroaki OHYA-NISHIGUCHI	35
Key words : Electron spin resonance, Spatiotemporal measurement, Nitroxide radical, Kinetics, In vivo
Abstract : Electron spin resonance (ESR) spatiotemporal measurements are reviewed in this paper. ESR spatiotemporal analysis has been developed to investigate the dynamics of free radicals in vivo. The core techniques of the analysis are the developments of a rapid field-scan in vivo ESR spectrometer and the imaging system, which have been constructed by introducing the microwave frequency lower than 1 GHz, and new resonators such as loop-gap type and surface-coil type. Two methods have been applied for rats or mice receiving the administration of free radical probes: One is the method that the kinetic constant of the probe at a target position is determined from the time course of the gray scale value in a series of ESR images. Second is that the ESR signal of the probe is directly observed by placing a small sensor at a target position. For typical example, the rate constants of first-order kinetics in rat's brain were determined for the probe which could enter to the brain through blood-brain barrier.

[-> return]

title	author	page
MEG and Imaging of Brain Functions	Tsunehiro TAKEDA, Akira UCHIDA, Hiroshi ENDO	43
Key words : MEG, Brain, Vision, Visualization, Accommodation
Abstract : Whole head type MEG (MagnetoEncephaloGraphy) is most promising for the study of human higher brain functions, because it has high temporal resolution and reasonable spatial resolution. This paper explains the basic of MEG measurement, some recent applications and the study for the visualization of the measured MEG data.

[-> return]

title	author	page
Brain Imaging and Neurosurgical Operation	Takamasa KAYAMA	51
Key words : Neurosurgery, Brain imaging, MEG, Gyrectomy
Abstract : It is well known that neurosurgical treatments are completely different comparing with other medical fields. Brain has functional localization. We, neurosurgeon should preserve brain function, ex motor, sensory,visual and speech function when we operate brain lesions. Recent progress of brain imaging is so much. I mentioned an advantage of brain imaging to neurosurgical operation. NEG disclosed motor area,sensory area, visual area on each cases. These informations much contributed to decrease postoperative complications. I believe in future neurosurgeons operate with computer aided surgery using informations detected by brain functional imaging.

< Papers >

[-> return]

title	author	page
Real Time Radiology Using Intersocket Communication	Yasutomi KINOSADA·Akihiro TAKADA·Tomoho MAEDA, Masato HORINO· Shin-ichiro NISHIDA·Norihisa MATSUI·Minoru HOSOBA and Mayumi KITANO	56
Key words : Telemedicine, Teleradiology, Real time radiology, Computer Workstation, CRT diagnosis
Abstract : Teleradiology has been a useful system to perform remote diagnosis using CRT units. However, a conventional teleradiology system can not show the advantage of the ∪real time∩ nature of the system clearly, because most systems have been dedicated to CRT diagnosing of images obtained at and transmitted from a remote hospital. The study, therefore, is to introduce the new two concepts for teleradiology : ∪real time radiology∩ and ∪seamless operation of a workstation to view images from both local and remote hospitals∩ into a teleradiology and to establish such a system. ∪Real time radiology∩ is a system, which can cover diagnostic practices in real time by radiological specialists at different locations. We designed a real time radiology system with a multimedia diagnostic workstation, which was established under the new teleradiology concepts and could be used for cooperative diagnosis by interchanging radiological images bilaterally in real-time during an audio-visual discussion. An intersocket communication technology was used to establish such a new real time radiology system. In this article, we show the system configuration and present the results of evaluation from technical and clinical points of view on real time radiology.

[-> return]

title	author	page
Reduction of MRI Artifact due to 3-D Rigid Motion	Reza Aghaeizadeh ZOROOFI, Yoshinobu SATO, Hiroaki Naito, Shinichi Tamura, Shinji Shimojo, Hironobu Nakamura	69
Key words : MRI, Rigid motion, Artifact correction, 3-D artifact, Non-uniform sampling reconstruction
Abstract : Patient motion plays a major role in the degradation of image quality provided with magnetic resonance imaging (MRI). In this work, we present a technique to suppress MRI artifact arising from 3-D rigid motion. Toward this goal, a multiple 2-D slice technique has been employed to acquire the MR image-data. The collected MRI signal corresponding to the object movement is affected by the non-uniform sampling and phase errors. With given motion parameters and an assumption of only inter-view effects for the 3-D motion, a reconstruction algorithm based on using the multi-slice data in a bilinear interpolation and superposition method is introduced to reduce MRI artifact. To estimate unknown motion parameters, a minimum energy method is applied which utilizes the fact that a 3-D rigid motion increases the measured energy outside the region of interest (ROI) associated with each image of a multi-slice data acquisition technique.

[-> return]

title	author	page
Detection of Pulmonary Nodules on Lung X-ray Images: Studies on multi-resolutional filter and energy subtraction images	Akira SAWADA, Yoshinobu SATO, Shoji KIDO and Shinichi TAMURA	81
Key words : Conventional lung X-ray images, Energy subtraction X-ray images, Pulmonary nodules, Laplacian of gaussian filter, Minimum directional difference filter, Multi-resolutional filter, Performance evaluation
Abstract : The purpose of this work is to prove the effectiveness of an energy subtraction image for the detection of pulmonary nodules and the effectiveness of multi-resolutional filter on an energy subtraction image to detect pulmonary nodules. Also we study influential factors to the accuracy of detection of pulmonary nodules from viewpoints of types of images, types of digital filters and types of evaluation methods. As one type of images, we select an energy subtraction image, which removes bones such as ribs from the conventional X-ray image by utilizing the difference of X-ray absorption ratios at different energy between bones and soft tissue. Ribs and vessels are major causes of CAD errors in detection of pulmonary nodules and many researches have tried to solve this problem. So we select conventional X-ray images and energy subtraction X-ray images as types of images, and at the same time select ⑧2G (Laplacian of Guassian) filter, Min-DD (Minimum Directional Difference) filter and our multi-resolutional filter as types of digital filters. Also we select two evaluation methods and prove the effectiveness of an energy subtraction image, the effectiveness of Min-DD filter on a conventional X-ray image and the effectiveness of multi-resolutional filter on an energy subtraction image.

March(Vol.17, No.2)

[-> return]

title	author	page
Flat Panel Sensors: Questions and Answers	Guy FRIJA	99
Key words : Digital X-Ray, Computed radiology
Abstract : After years of R&D works, digital X-Ray detectors developers appear to be ready for starting the commercial step of flat panel sensors. By comparison with conventional film-screen, the dynamic of digitized images is higher and should result in enhanced diagnosis. Finally, digitized images are especially subject to protocols for procedures and to sophisticated post processing developments including automatic detection of abnormalities (for example, breast microcalcifications). There is a wide range of digital X-Ray technologies from different companies. It seems that the cost of a large FPD would be 150,000-200,000 USD, making a total X-Ray department digitization prohibitive, unless prooven significant return on investment attributable to film, and chemicals replacements, but also to patient throuput enhancement and to better staff's productivity can be demonstrated. It is expected that first commercialization will be developed with chest dedicated systems, followed quickly by full field digital mammographic systems and by small size (9 inches) image intensifiers. The inability to use FPD for portable X-Ray is one of their main limitations. Therefore it is not sure that FPD will render CR phosphore plates technology obsolete.

[-> return]

title	author	page
A Direct Digital Image Capture System: The Future of Digital X-ray	Jeffrey H. BELL	105
Key words : DirectRay, Digital X-ray, Direct image capture
Abstract : Conventional projection radiography accounts for the largesti volume of medical images produced in a radiology department, yet it is the last area to enter the electronic age. The direct digital capture of X-rays has become a reality through recent technological developments in the electronic and semiconductor industries. The conventional analog and cutting-edge digital modalities currently available are considered either indirect or direct based on≥respectively, their use of or elimination of light. Of those available, DirectRay(TM), a singular direct-to-digital technology, offers image quality, along with clinical test results and market strategy. To better understand what this means to medical imaging, let's briefly review four items: ˇ the evolution of digital imaging technology; ˇ the different types of digital image capture systems; ˇ the science behind DirectRay technology; ˇ and regarding this new technology-the results to date, its future and the needs of radiologists.

[-> return]

title	author	page
Development of a Direct-conversion-type Flat-panel Detector for Radiography and Fluoroscopy	Shin-ichi YAMADA	110
Key words : Flat-Panel detector, Dynamic Images, Direct conversion, High resolution
Abstract : This paper describes the X-ray flat-panel detector technology which will be able to take the place of all conventional X-ray detector technologies. Hospitals are rapidly changing from handling information with analogue systems to using digital systems, accompanying the advances being made in this digital information age. The X-ray flat-panel detector will convert 2D distribution of X-ray intensity to digital image data instantly. Until now, there has not been an X-ray flat-panel detector that can completely take the place of analogue X-ray films with screens (S-F) and can acquire high-speed dynamic X-ray images with high image quality, but we have realized this by developing a direct-conversion-type flat-panel detector. The latest research result confirmed the increase of X-ray absorption capability, spatial resolution which exceeded a regular type of S-F and images without distortion in a large field of view. This detector is a compact size, just like an X-ray cassette, and can be used for both radiography and fluoroscopy, which can take the place not only of S-F but also Imaging Intensifier (I.I.) -TV. The X-ray flat-panel detector will cause dramatic changes to all X-ray systems and we can say specially that the direct-type flat-panel detector has proven to be a step toward the coming era of real full-digital hospitals.

[-> return]

title	author	page
Development of Flat Panel Digital Radiography System	Tatsuya YAMAZAKI, Masakazu MORISHITA, Noriyuki KAIFU, Yutaka ENDO	117
Key words : Radiography, Digital, Flat panel detector, Amorphous silicon
Abstract : We developed the Digital Radiography System CXDI-11 which digitizes the X-ray image in high quality by using a self-developed flat panel detector. The CXDI-11 has a large image area of 43cm∵43cm (17"∵17"), and it can display the image on the pre-view monitor after only 3 seconds of exposure. In this report, we present the principle and the physical characteristics of the CXDI-11. The X-ray detector installed in the CXDI-11 is a combination of a rare-earth scintillator and an amorphous silicon flat panel detector (LANMIT). The X-ray is converted to the visible fluorescent light at the scintillator and the light is detected by the LANMIT. The image-processed data is transferred to the DICOM 3.0 conformed devices such as the diagnosis work station, the archiver and the laser imager through the network. We also show some measurement results of the dynamic range, the pre-sampling Modulation Transfer Function and the tube voltage dependent sensitivity. The CXDI-11 is superior in real time operation and image quality, thus it is the digital radiography system of the next generation.

[-> return]

title	author	page
Integrated Digital Radiography with a Flat-panel Sensor	Ulrich NEITZEL	123
Key words : Digital radiography, Flat-panel sensor
Abstract : In this article, a short review of digital radiography systems available today is followed by a description of the emerging technology of electronic flat-panel sensors. These sensors allow fully electronic acquisition of all standard radiographic images with excellent image quality and facilitate the smooth integration of radiographic examinations into a digital radiology department. Specifications of these sensor types are reported and their operational advantages are discussed.

[-> return]

title	author	page
Integrated FD-Systems in a Networking Environment	Reiner F. SCHULZ	130
Key words : Radiography, Solid state X-ray detectors, Amorphous silicon
Abstract : Systems for Radiography equipped with solid state X-ray detectors are being introduced into the market. Users can benefit most from the new technology, if the X-ray detector is designed with regard to application and system aspects, and the system is embedded in a networking environment.

[-> return]

title	author	page
Surgery of Facial Cancer and Face Appearance Estimation	Eiji TAKAHASHI, Toyohisa KANEKO, Koichi HIROTA, Kiyotaka UCHIYAMA, Ryuzo SEKIGUCHI and Noriyuki MORIYAMA	137
Key words : Surgical Simulation, Facial Cancer, Bubble Meshing, Face Appearance Estimation
Abstract : Our objective is to build a system for predicting a cancer patient's facial appearance after surgery. We need facial 3D model with skin color for our objective.In order to color the 3D polygonal surface model extracted from CT data, we employed patient's three color photographs : the front view and two side views. We developed a technique to place the photographs exactly on the colorless 3D model based upon manually selected landmarks on the face. A temporal change in the facial appearance after surgery is of interest. We employed a spring-mass network model for soft tissue simulation, which requires nodes distributed as uniformly as possible. We used a physical approach to placing bubbles in a 3D space, where bubbles attract or detract each other, depending upon mutual distances. Special attention was paid on the surface where bubbles were allowed to stay on a limited number of polygonal veretices. As a deforming example, we simulated surgey of upper jaw cancer. At first, we removed a part of upper jaw bone and then the corresponding soft tissue is stiffened.

[-> return]

title	author	page
Discriminating Spinal Deformity Employing Two Characteristic Axes on the Moire Images of Human Backs	HyoungSeop KIM, Seiji ISHIKAWA, Kiyoshi KATO and Yoshinori OHTSUKA	148
Key words : Symmetry, Potential symmetry, Principal axis, Moire image, Spinal deformity
Abstract : Spinal deformity is a serious problem mainly for teenagers and medical doctors inspect moire topographic images of their backs visually for the screening. If a subject is normal, the moire image is almost symmetric with respect to the middle line of the subject's back, otherwise it shows asymmetric shape. In this paper, we propose a technique for discriminating spinal deformity employing the angle made by two characteristic axes, i.e, the principal axis and the potential symmetry axis. The two angles measured at two areas on a subject's back compose a 2-D feature space and inspected cases are devided into two clusters in the space by a linear discriminant function based on the Mahalanobis distance. Given 120 cases, 60 normal and 60 abnormal, the leave-out method was applied to the classification experiment and 69.2% of the classification rate was achieved with respect to employed test data.

[-> return]

title	author	page
A 3D Optical Head Motion Measurement System and Its Primary Application in Stereotactic Radiosurgery	Jiang KAI, Hiroya SHIOMI, Toshihiko SASAMA, Yoshinobu SATO, Takehiro INOUE, Shinichi TAMURA, Toshihiko INOUE	155
Key words : Head motion, Optical measurement, Radiosurgery, Precision
Abstract : Head motion is a significant source of therapy-effect degradation in stereotactic radiosurgery (SRS). In this study, a 3D optical head motion measurement system with a precision less than 0.1mm has been developed by our group. During the clinical SRS treatment experiments, three dimensional (3D) positions of infrared light emission diode (ILED) markers on the head - chin, nose tip, forehead and ear were measured with a data sampling rate of once per second for 30 minutes, in frame and frameless sessions, respectively, by this system. The t-test and linear correlation analysis have been used, and the results indicate that the difference of head motion with moldable thermoplastic frame on and without frame was highly statistically significant (p<0.01), and the fixation of the moldable thermoplastic frame in X axial direction (defined in local treatment-couch coordinate system) needs to be reinforced to realize better fixation effect for high degree of treatment precision. It is also shown that the head motion was mainly due to the rotation about the occiput-couch contact point, however, the chin is the most easily movable area on the head. The result implys that in the newly developing frameless SRS treatment, the landmarkers should be attached to the relatively static areas such as forehead and ear.

[-> return]

title	author	page
3D Heart Imaging System Using Echocardiogram and a Digitizer for a US Probe Location	Osamu OSHIRO, Masayuki NAMBU, Ayumu MATANI and Kunihiro CHIHARA	165
Key words : Ultrasound B-mode image, 3D image, ECG, Digitizer, VRML, Tele-medicine
Abstract : This paper describes a three-dimensional(3D) heart imaging system, where a ultrasound(US) B-mode image triggered with electrocardiogram(ECG) was obtained with a conventional US diagnostic device and the location and orientation of a US probe were detected with a 3D digitizer. The B-mode image was projected onto a virtual 3D space and then a 3D image in the format of virtual reality modeling language(VRML) was presented on a CRT. We performed an experiment to reconstruct a 3D heart image on systole using this system, which indicates that the system has the potential for the diagnosis of heart disease. Furthermore, a 3D image was displayed with a world wide web(WWW) browser and a VRML plug-in, which demonstrates the possibility of this system for tele-medicine.

[-> return]

title	author	page
Inner Volume RARE with Navigator Echoes for Interventional MR Imaging	Kenji SHIMIZU, Koichi OSHIO, Ron KIKINIS, Ferenc A JOLESZ, Shinichi TAMURA, Junpei IKEZOE	173
Key words : Device tracking, Rapid imaging, Inner volume, Rapid acquisition with relaxation enhancement (RATE), Local Look (LoLo), Interventional MR imaging
Abstract : We propose a new imaging technique based on an inner volume modification to rapid acquisition with relaxation enhancement (RARE) to dynamically monitor interventional procedures within MR imaging. With use of the projection navigator echoes, the invasive device such as the biopsy needle can be rapidly tracked. This technique was implemented on a standard MR imager and its feasibility was demonstrated in phantom studies. Our approach can be readily incorporated into the enhanced version of inner volume RARE sequence.

May(Vol.17, No.3)

[-> return]

title	author	page
Images of FISH(fluorescence in situ hybridization) and Its Applications for Human Molecular Cytogenetics	Johji INAZAWA, Yoji FUKUDA	185-190
Key words : FISH, SKY, CGH, Chromosome Aberrations, Cancer
Abstract : FISH is a powerful technique for direct mapping and ordering of genes or anonymous DNA fragments. Recent advances in FISH technique have provided useful tools such as spectral karyotyping (SKY), M【FISH, and CGH. Images obtained by these techniques also confer valuable information for positional cloning genes responsible for hereditary disorders and cancer.

[-> return]

title	author	page
CGH∈Comparative Genomic Hybridization∷Analysis	Kohsuke SASAKI	191【196
Key words : FISH, Cancer, Cytogenetics, DNA, Chromosome
Abstract : The present brief paper introduced a technique of CGH(comparative genomic hybridization) which provides information concerning DNA sequence copy number losses and gains in a single hybridization. CGH allows molecular cytogenetic analysis of solid tumors. This has advantages over conventinal cytogenetic methods, because neither cultures of tumor cells nor preparations of metaphase spreads are required for cytogenetic analysis. A CGH analysis system is an image analyzer consisting of a fluorescence microscope, CCD camera and computer. In CGH analysis, DNA samples taken from tumor and normal tissues which are labeled with different fluorochromes are simultaneously hybridized to normal metaphase spreads, and intensity ratio of these two fluorescence signals represents the changes in DNA copy number of each chromosome or chromosomal segment.

[-> return]

title	author	page
Imaging of Laser Scanning Cytometer	Keiji KAWAMOTO, Yuji TSUKAZAKI, Yoshihiro NUMA, Shiguang ZHAO	197【202
Key words : Laser scanning cytometer, Imaging, Flow cytometry, Brain tumor, Cell kinetics
Abstract : dLSC was designed for detecting the single cells on the slide glass by rapid scanning with the laser beam. This machine shows the same ability of flow cytometry for cell kinetics and is also possible to observe the cell morphology. Imaging of LSC shows the DNA histogram with use of PI staining and cell morphology at the same time by recall operating system of selected cells from the histogram. LSC could be applied to the analysis of chromosome, FISH and pathological diagnosis of biopsy or smear.

[-> return]

title	author	page
Fifty Years of Attempts to Automate Screening for Cervical Cancer	Ewert BENGTSSON	203【210
Key words : cervical smear prescreening, Pap-smear automation, digital image analysis, automated cytology, pattern recognition
Abstract : Cervical cancer is one of the most deadly and common forms of cancer among women if no action is taken to prevent it, yet it is preventable through a simple screening test the so called Pap-smear. Each year at least a hundred million such samples are collected and screened and the numbers are increasing. This creates a tedious mass inspection task. Numerous attempts have been made over the last 50 years to create systems that automate this task. The author has been involved in research and development of both automated and interactive cell analysis systems during the last 25 years. Based on this experience and studies of the literature and relevant WWW-pages the developments of the field are traced through the years and some comments are made about how the different generations of systems relate to each other. In spite of all the automation efforts, still no generally accepted automated prescreening device exists on the market. The main reason for this failure is the great pattern recognition capabilities needed to distinguish between cancer cells and all other kinds of objects found in the specimens: cellular clusters, debris, degenerate cells etc. Improved algorithms, the ever-increasing processing power of computers and progress in biochemical specimen preparation techniques make it likely that eventually useful automated prescreening systems will become available.

[-> return]

title	author	page
Observation of Living Cells with X-ray Microscopy	Hideaki SHIMIZU, Toshikazu MAJIMA, Toshihisa TOMIE	211 -216
Key words : X-ray contact microscopy, Laser-plasma x-ray source, Atomic-force microscopy,Internal structures of living cells
Abstract : X-ray microscopy is promised to be a useful tool for studying living specimen. The construction of a flash and contact type X-ray microscope is reported and the X-ray images of living cells are demonstrated. Internal structures of the cells were clear, and the carbon density was estimated from the images.

< Papers >

[-> return]

title	author	page
Reduction of False -Positive Shadows in Computer-Aided Diagnosis System for Chest X-ray CT Images	Atsushi SUGIYAMA·Satoru KAMANO·Shinji YAMAMOTO·Mitsuomi MATSUMOTO·Yukio TATENO·Takeshi IINUMA·Tohru MATSUMOTO	217-227
Key words : X-ray CT images, Feature analysis, Computer aided diagnosis, Lung cancer, Image processing
Abstract : In this paper, we report how we have improved our method for reducing the false-positive rate in computer-aided diagnos is in the Lung cancer Screening system by CT(LSCT). We had changed the target size for the detection of cancers from more than 10mm to more than 5mm in diameter. As a result of this change, many false positives with diameter of less than 10mm were observed. In order to reduce this false-positive rate, we propose a method involving feature analysis of can didate cancers using feature parameters. First, we calculate 13 feature parameters containing 4 new parameters that we have defined. These new parameters were designed to reflect the shape features of candidate cancers less than 10mm in diameter; specifically, revised circularity, modified elongation, convexity, and the ratio of convergent figures. Next, candidate cancers were divided into 4 classes based on the area and the thickness parameter, and cancers and false positives were distinguished based on the values of the feature parameters in each class. This reduction method was applied to LSCT images of 83 samples, and the method reduced the total number of candidate cancers from 12399 to 977. The average number of false positives per sample was 10.7,without any additional false positives. These results show the effec tiveness of the proposed method.

[-> return]

title	author	page
Detection and Segmentation of Intersection and Branch Structures in Medical Images Using Orientation Space	Jian CHEN, Yoshinobu SATO·Shinichi TAMURA	229-244
Key words : Line segmentation, Junction, Gabor filter, Orientation space, Vessel segmentation
Abstract : The purpose of this paper is to present appropriate line segmentation for intersections (X-junctions) and branches (T-junctions). In the local regions of intersections and branches, multiple orientations occur. We propose a novel representation called the "orientation space", which is derived by adding the orientation axis to the abscissa and the ordinate of the image. The orientation space representation is constructed by treating the orientation parameter, to which Gabor filters can be tuned, as a continuous variable. The problem of multiple orientation line segmentation is dealt with by thresholding 3D images of the orientation space and then detecting the connected components therein. In this way, X-junctions and T-junctions can be separated effectively. Experimental results obtained using synthesized and real biomedical images are presented. In particular, overlapping vessels in an X-ray coronary angiogram were well segmented by orientation space filtering.

[-> return]

title	author	page
Application of Constrained Optimization Using Augmented Lagrange Functions to Photon Attenuation Correction in SPECT	Shigehito SUZUKI, Misato WAKABAYASHI, Susumu KUWAMURA, Keiichi OKUYAMA	245-260
Key words : Attenuation correction, SPECT, Nonlinearly constrained optimization, Augmented Lagrange function, Conjugate gradient method
Abstract : The attenuation correction method in SPECT based on the nonlinear optimization which we previously reported (Med Imag Tech 16: 237-251, 1998) is extended to include a non-negativity constraint on pixel values. This optimization is conducted by minimizing augmented Lagrange functions using the conjugate gradient method, together with the quadratic function method of the step-width algorithm. Further, smoothing for search direction data is included in the present method to reduce statistical noise in the reconstructed image. The constrained optimization did not make all pixel values in the im age non-negative, although the negative values were fairly close to zero. It nevertheless provided improvement in image noise: it prevented the rise in image noise with increasing iterations because the constraint restricted the distribution of pixel values to a narrow interval. The optimization produced numerical phantom images similar to the original. Clinical application also demonstrated that the proposed method is effective.

[-> return]

title	author	page
Enhancement of Microcalcifications in Mammograms Using Dyadic Wavelet Analysis	Yoshikazu UCHIYAMA, Koji YAMAMOTO	261-271
Key words : Wavelet, Contrast enhancement, Mammogram
Abstract : Experience has shown that the contrast of an image can be enhanced by increasing the magnitude of the edge gradient. Wavelet transformation provides multi-scale edge gradient images. This may make it possible for fine structures of a certain size to be amplified by modifying the modulus of the gradient images with suitable weights. In this paper, we propose a method for determining the optimal weights to enhance microcalcifications in mammographic images. By applying our method to mammographic images, we have demonstrated its ability to selectively enhance microcalcifications.

[-> return]

title	author	page
Automatic Extraction of Casts in Urinary Sediment Images	Ning-Feng ZENG, Keiji TANIGUCHI, Hong ZHU, Sadakazu WATANABE, Yutaka NAKANO, Hiroyuki NAKAMOTO	273-283
Key words : Extraction of casts , Urinary sediment image, Iterative scheme, One- dimensional average, Cast symbolic description, Density threshold
Abstract : The extraction of casts from urinary sediment images is a difficult task, because the difference in gray level between the casts and their background is almost zero. A special example of this difficulty is the hyaline cast, which is transparent. In this paper, we describe a novel approach to extract casts in urinary sediment images. We introduce an iterative scheme of one-dimensional averaging of the gray level to calculate the exact average gray level of the background. Based on these values, we are able to convert the original image into a binary image. Furthermore, to allow the computer to identify casts in the binary image, we summed up the features of casts from a large number of typical samples, which were defined as a set of symbolic descriptions of casts. This set of symbolic descriptions is used to determine whether or not an the element in the binary image is a cast. From the experimental results for 120 urinary images, we conclude that this method is effective in cast extraction. The rate of successful identification is up to 80 percent.

September(Vol.17, No.5)

[-> return]

title	author	page
Non-Invasive Imaging of Absorption Changes in Rat Brain by NIR Transillumination	Yoshinori TAKA, Kaoru SAKATANI, Yuji KATO and Koichi SHIMIZU	545-555
Key words : Near-infrared light, Cerebral function, Transillumination, Non-invasive imaging
Abstract : A transillumination technique has been developed for the non-invasive imaging of cerebral function with near-infrared(NIR) light. The feasibility of such imaging has been demonstrated in theoretical analysis. Even in a scattering medium such as mammalian tissue, we can obtain the two-dimensional distribution of absorbance changes in the scattering object. In an experiment, it was found that changes in blood volume and oxygenation could be detected in the transillumination images. Further, the possibility of detecting the localized changes in the brain caused by sensory stimulation was explored. The local increase in blood volume in the cerebral somatosensory area was successfully detected in the transillumination image of the brain when electrical stimulation was applied to the rat forelimb. In an experiment involving the simultaneous measurement of somatosensory evoked potentials, it was confirmed that this increase in blood volume corresponded to the increase in cerebral activity. The result of these experimental studies demonstrate the feasibility of functional imaging by NIR transillumination.

[-> return]

title	author	page
Development of Correction Algorithm for Motion Artifacts in MRI Real-time Processing	Masahiro TAKIZAWA, Tetsuhiko TAKAHASHI, Kenji TAKIGUCHI and Hiromichi SHIMIZU	556-562
Key words : MRI, Motion correction, Phase, Navigator echo
Abstract : In MRI, navigator correction is often performed in order to reduce artifacts that are generated when the subject moves. We have developed a highly accurate and stable algorithm that is appropriate for real-time processing. Since this algorithm obtains the motion information for the subject directly from the phase shift of the echoes and requires no iterative processes, it is suitable for high-speed processing. Experiments using a multi-shot EPI sequence showed that if the subject moves only within the range from -10 to 10 mm during the execution of a pulse sequence, correction is performed well and motion artifacts are almost completely eliminated.

[-> return]

title	author	page
Visualization of the 3D Shape of the Articular Cartilage of the Femoral Head from MR Images	Tetsuya KUBOTA, Yoshinobu SATO, Katsuyuki NAKANISHI, Hisashi TANAKA, Nobuhiko SUGANO, Takashi NISHII, Kenji OHZONO, Takahiro OCHI, Hironobu NAKAMURA and Shinichi TAMURA	563-575
Key words : MR slice images, Articular cartilage of femoral head, 3D shape reconstruction, 3D Multiscale sheet filter,Volume rendering
Abstract : This paper describes methods for visualizing the three-dimensional (3D) cartilage thickness distribution from MR images. Cartilage thickness is one of the most important factors in joint diseases. Although the evaluation of cartilage thickness has received considerable attention from orthopedic surgeons and radiologists, evaluation is usually performed based on visual analysis or measurements obtained using calipers on original MR images. Our aim is to employ computerized quantification of MR images for the evaluation of the cartilage thickness of the femoral head. First, we extract an ROI and interpolate all ROI images by sinc interpolation. Next, we extract cartilage regions from MR images using a 3D multiscale sheet filter. Finally, we reconstruct 3D shapes by summing the extracted cartilage regions. We investigate partial volume effects in this method using synthesized images, and show results for in vitro and in vivo MR images.

[-> return]

title	author	page
Development of a Mass-Classification System for Mammograms	Takeshi HARA, Yoshinobu TANI·Hiroshi FUJITA and Tokiko ENDO	577-584
Key words : Mammographic mass, Classification, Fractal, Computer-aided diagnosis
Abstract : We have developed a mass-classification system using fractal dimensions and texture analysis to classify masses detected on mammograms. This mass-classification method calculates each fractal dimension from the contour patterns that are extracted from binarized images. We were able to analyze the shapes of masses without subjective contour information. We classified masses as benign or malignant by using an artificial neural network (ANN) which was used to analyze the normalized mean and slope values of the fractal dimensions, texture feature values, and the degree malignant features by a pendulum filter to detect spicules. The results for a database of 25 cases that were diagnosed by a radiologist showed a sensitivity 85.7% and a specificity of 81.8%. We conclude that our method is effective for discriminating between various types of mammographic masses.

[-> return]

title	author	page
Display System for Image Plane Position and Orientation in Ultrasonographic Examination of the Heart	Yoshikazu NAKAIJMA·Yoshinobu SATO·Hiroaki NAITO·Yuji OGATA·Shinichi TAMURA·Takehiko MORIGUCHI and Jun ARISAWA	585-592
Key words : Diagnostic assistance, Image guidance, Medical navigation, Ultrasonographic examination, Heart
Abstract : In ultrasonographic examinations,the image plane can be positioned at any location or In any orientation in space or within the target organ. Thus, during ultrasonographic examinations, medical staff should clealy grasp the position and orientation of the target organ relative to the ultrasonographic imaging plane.Particularly in the examination of the heart, it is difficult to determine the best position and orientation of the ultrasonic probe to acquire good images. The reasons are as follows: (1) the position and orientation are restricted by the ribs, and (2) heart size and position differ from person to person. For these reasons, we have developed an image guidance system for the ultrasonographic examination of the heart. Our system shows the positional relationships among an X-ray image, a heart model, and the ultrason ographic imaging plane, and is useful for medical training and for clinical applications. In this report, we describe the performance of our system when applied to the examination of a healthy subject.

[-> return]

title	author	page
Locating Magnifying Windows for Observation of Gland Tubule Structures in Stomach Tissue Images	Yutaka NAKANO·Keiji TANIGUCHI·Sadakazu WATANABE and Tatsuya NAKAMOTO	593-600
Key words : Stomach tissue image, Gland tubule structure, Spatial clustering
Abstract : It is often the case that the gland tubule regions extracted from stomach tissue images should be observed in detail of higher magnification. This paper discusses two methods for locating the magnifying windows on a whole-tissue image. The first method is based on the spatial clustering where gland tubule regions are concentrated when the cluster size is less than the window's diameter. In the second method, which is designed to be economical in terms of the number of windows, each window's location is selected sequentially from among feasible ones based on combinatorial relationships. In our experiments using actual samples with various appearances, both methods gave good locations, and the number of windows in the second method was less than that in the first.

[-> return]

title	author	page
Procedure for Extracting Vessel Shadows in Chest X-ray Images Using a Deformable Model	Chae-uk SONG, Akinobu SHIMIZU, Jun-ichi HASEGAWA, Jun-ichiro TORIWAKI and Masaki MORI	601-613
Key words : Chest X-ray images, Pulmonary emphysema, Extraction of blood vessel shadows, Deformable model, Scale space
Abstract : To quantify the degree of progression of pulmonary emphysema, which is considered to be strongly correlated with the width of blood vessels, we propose a procedure for automated extraction of peripheral blood vessel shadows in conventional chest X-ray images. This procedure mainly employs 1) a deformable model that has been designed based on information concerning the shapes of blood vessels and is expected to be useful in extracting blood vessels, especially those located in backgroud regions with a low signal-to-noise ratio, and 2) a scale space scheme that is useful even if there is little advance information concerning the sizes and positions of blood vessels. Furthermore, we add to this procedure the ability to set up an initial model automatically and to modify the model's topology during deformation. We apply this procedure to actual chest X-ray images, evaluate its performance experimentally, and show its superiority to the conventional difference filtering method.

November(Vol.17, No.6)

[-> return]

title	author	page
Contrast Enhanced MR angiography	Makoto AMANUMA	633-638
Key words : MR angiography, Contrast media, Three dimensional imaging
Abstract : Contrast enhanced MR angiography (MRA) is the technique based on combination of rapid 3D imaging and bolus injection of contrast media. This technique has now widely accepted and used in routine clinical works for imaging vascular structures of the body and lower extremities. With many recent technical developments including an injection-imaging timing detection, data subtraction, and zero-filled interpolation, this method can now provide more sophisticated vascular images with a high spatial resolution and also, temporal resolution. Recent introduction of table-moving technique has made MRA covering larger part of the body possible. MR digital subtraction angiography (MRDSA) using a rapid 2D acquisition suggests clinical potential of the high temporal resolution DSA-like images. The selection of these various techniques depends on the imaging purpose, hardware capacity, and availability of the other imaging modalities. In this article these several new techniques with which contrast enhanced MRA has developed and recent clinical applications are briefly reviewed.

[-> return]

title	author	page
Quantification of MR Spectroscopy --Possibility for clinical application and subject for future development--	Masafumi HARADA, Hiromu NISHITANI	639-644
Key words : MRS, Quantification, Proton
Abstract : This review paper introduces the recent tendency of development for MR spectroscopy. The main target is now proton MRS and the tendency is classified into three directions such as:1)Automatic procedure 2)Quantification 3) Reduction of time consuming The improvement of shimming quality is much contributed to the automation for proton MRS measurement. The internal water reference method for quantification of proton MRS is recently adopted widely and the reports showing utility of quantified MRS is now increasing. We showed one example of pediatric CNS diseases in which quantification of proton MRS was useful for differential diagnosis. The reduction of the measurement time is tried to automate measurement procedure, but still difficult when long TR sequence is used. To reduce measurement time, the improvement of sequence such as fast SE would be indispensable. MRS is one of attractive noninvasive method to observe metabolism, but now still progressing and developing for clinical utility.

[-> return]

title	author	page
Review of Cardiovascular MRI: Clinical Results and further developments	Osamu TAKIZAWA, Daisy CHIEN, Gerhard LAUB	645-651
Key words : Cardiovascular MR, Coronary angiography, Myocardial perfusion, Dedicated cardiac MR system
Abstract : Magnetic resonance (MR) has many intrinsic advantages for cardiovascular applications: it offers excellent soft tissue contrast and has the ability to acquire three-dimensional and/or oblique images. Furthermore, one can non-invasively view the proximal branches of the coronary arteries, evaluate myocardial perfusion, and quantify blood flow. With the new myocardial viability technique, one can even evaluate the size of an infarct in around 10 seconds.Dedicated cardiovascular MR has emerged recently from the research arena into clinical reality. A key feature of these dedicated systems is that they offer much faster techniques than ever before. The new techniques are expected to further expand the clinical utility of MR for diagnosing cardiovascular diseases.

[-> return]

title	author	page
Brain Functional MR Imaging	Hiroki YAMADA	652-657
Key words : Functional MR imaging, Brain activation, Echo planar imaging
Abstract : Functional magnetic resonance Imaging (fMRI), which is a non invasive imaging tool for detecting neuronal activity induced by the external stimuli, has recently become available for clinical examinations of the brain function during development. Blood oxygenation level dependent (BOLD) contrast is thought to be the contrast mechanism of fMRI especially in clinically available 1.5 T MRI system. In this article, clinical utility of the fMRI was presented for assessing brain function in patient with brain tumor, ischemic or traumatic brain damage, and for detecting dynamic metabolic changes during the brain maturation.

[-> return]

title	author	page
Interventional Magnetic Resonance Imaging: Concepts and Applications in Neuroradiology	Michael WENDT, Jeffrey L. DUERK, Jonathan S. LEWIN	658-667
Key words : Interventional MRI, Fast imaging, Intraoperative MRI, Minimally invasive therapy
Abstract : Recent trends in minimally-invasive methods for diagnosis and treatment have created a burgeoning interest in the use of MR imaging for guidance of radiological and surgical procedures. This report discusses the hardware and software developments in the last few years that have made interventional and intraoperative MR imaging a reality. These advances include new concepts in MR system design, along with supplemental technological developments in MR pulse sequences, guidance systems, and in-room monitors. This report also describes our experience in using MRI to guide minimally invasive therapies, such as biopsies and radio frequency (RF) ablation of tumors in the liver, as well as to monitor neurosurgical procedures in a fully functional surgical environment.

[-> return]

title	author	page
Extraction of Renal Glomerular Regions Using a Non-linear Difference Filter and Region-estimation Model	Hong ZHU, Keiji TANIGUCHI, Sadakazu WATANABE, Yutaka NAKANO, XiaoMang ZHANG	669-680
Key words : Glomerular region, Bowman's capsule, Non-linear difference filter, Model for segmentation, Boundary estimation
Abstract : It is very important to extract precisely the glomerular region from a kidney image because important information reflecting kidney disease may be contained in the glomeruli. The boundary of the glomerular region is sometimes non-closed and not clear. It is difficult to obtain the glomerular region with high precision by applying conventional image processing techniques when the non-closed boundary is large. In this paper, we propose a new method to solve this problem. In the first step, a non-linear difference filter, which is used to enhance the Bowman's capsule region and to suppress areas both inside and outside of the glomerular region, is applied to binarize the region of the renal tissue image. In the second step, the non-closed boundary of the glomerular region is estimated by using a circular or elliptical model. In the final step, the glomerular region is easily extracted by this estimation. Of 150 samples tested using the present method, 142 were accurately processed.

[-> return]

title	author	page
Feature Extraction of Malignant Tumors on Mammograms	Masayuki MURAKAMI, Hidefumi KOBATAKE, Hideya TAKEO, Shigeru NAWANO	681-689
Key words :Mammography, CAD, Breast cancer, Tumor, Boundary characteristics
Abstract : It is widely known that important clues for discriminating between malignant tumors and benign tumors or normal tissues are found near boundaries. In this paper, a method for identifying the region adjacent to tumor boundaries using an iris filter is proposed. This filter is an adaptive filter applied to the gradient vector field. It is known that the optimal region of support of the iris filter theoretically coincides with the boundary of an ideally rounded convex region. Based on these characteristics, the iris filter was adopted to identify the regions adjacent to tumor boundaries on mammograms. Experimental results showed that the optimal region of support of the iris filter coincides well with tumor boundaries. Five feature parameters, which reflect the boundary characteristics, are also proposed. Experimental results showing the effectiveness of each feature parameter in discriminating between malignant tumors and other tumors or tissues are presented. A tumor detection system has been designed in which these five feature parameters and four other features are used to identify malignant tumors. Large-scale experiments to evaluate the performance of the proposed system have demonstrated the effectiveness of the proposed feature parameters.

[-> return]

title	author	page
Extraction of Fluorescent Dot Traces from a Scanning Laser Ophthalmoscope Image Sequence by Hough Transform on a Spatio-temporal Image	Shuzo YAMAMOTO, Yoshinobu SATO, Yoshikazu NAKAJIMA, Shinichi TAMURA, Hongfeng LIU, Seiyo HARINO	690-696
Key words :Ocular fundus image, Scanning laser ophthalmoscope, Spatio-temporal image, Fluorescent dot tracking, Hough transform
Abstract : The scanning laser ophthalmoscope allows the tracking of fluorescent dot motion,thereby enabling the flow velocities in the perimacular capillaries to be directly measured. These measurement can serve as an important index of local retinal soundness or reflect the whole-body circulation status in disorders such as diabetes. However, the fluorescent dots are so faint and unstable that it is difficult to detect them by conventional digital still image processing methods. To solve this problem, we generated a spatio-temporal image of the fluorescent dots in a capillary. We then applied the Hough transform to detect them as linear traces. By this method, 60% of the traces could be detected without error, which is sufficient to allow the calculation of mean flow velocity in a capillary.

[-> return]

title	author	page
Wired and Wireless Transmission System for Digital Images and Associated Diagnostic Reports	Tokuo UMEDA, Kiyofumi HANEDA, Tadashi KOYAMA, Hajime HARAUCHI, Kiyonari INAMURA, Junpei IKEZOE	697-703
Key words : Teleradiology, Radiological reporting system, Medical information system, Image transmission
Abstract : We have designed an economical and compact transmission system for teleradiology that can be used even in a disaster. Our system can transmit images and diagnostic reports using a wired LAN or a mobile wireless LAN. The system emplpoys conventional Ethernet LAN when using a wired LAN. On the other hand, the system employs transmission systems based on portable telephone lines, wireless LAN (Japan Radio Co., Ltd.) and Apple Remote Access (Apple Co., Ltd.) , when using a mobile wireless LAN. Transmission times for a 1-MB data package using a wired LAN, a mobile wireless LAN, and an ARA mobile wired LAN system were 5 seconds, 13 seconds, and 625 seconds, respectively. At 3 MB of per image, a wired LAN was 4 times faster than a mobile wireless LAN according to the results of our tests. The reason is that mobile wireless telephone lines sometimes suffer from poor transmission conditions, which must be compensated for by error correction functions. Chained structures of various types of image associated information worked well with adequate speed. However, important functions such as victim identification, as well as privacy and security concerns, must be evaluated by simulation of a disaster environment such as a major earthquake.

[-> return to top of index (1999)]