Search

In this study, we presented the forced flow and convective melting heat transfer characteristics of tetrabutylammonium bromide (TBAB) clathrate hydrate slurry (CHS) flowing through straight circular tubes. Pressure drops of TBAB CHS with the volume fraction in the range of 0–20.0 vol% were measured, and the influence of the volume fraction was analyzed. Power-law equation was used to describe the flow behavior of CHS, and the corresponding indicators including flow behavior index and fluid consistency coefficient indicated that TBAB CHS behaved as a pseudo plastic fluid. Based on the modified Reynolds number, it was found that the empirical correlation could be used to predict the flow friction factor of TBAB CHS. Heat transfer experiments were conducted to measure the local heat transfer coefficients of TBAB CHS flow under constant wall heat flux, which were found to be enhanced by the latent heat involved in the solid–liquid phase change. Local heat transfer coefficient decreased along the heating section until the volume fraction was smaller than a certain small value, then it increased in the following section due to the increase of Reynolds number. The effects of heating power, flow velocity, volume fraction on the local heat transfer coefficients were discussed. Correlations for predicting TBAB CHS heat transfer performance were proposed based on the experimental data in laminar, transitional and turbulent flow regions, respectively.

We have performed systematic studies of narrow Fe-K line (6.4 keV) flux variability and Ni-K line intensity for Seyfert galaxies, using {\em Suzaku} and {\em XMM-Newton} archival data. Significant Fe-K line variability of several tens percents was detected for a pair of observations separated by 1000--2000 days (Cen A, IC 4329 A, NGC 3516, and NGC 4151) and 158 days (NGC 3516). These timescales are larger by a factor of 10--100 than the inner radius of torus, consistent with the view that X-ray reflection by torus is a main origin of a narrow Fe-K line. Ni-K line was detected with $>2\sigma$ level for the Circinus galaxy, Cen A, MRK 3, NGC 4388, and NGC 4151. A mean and variance of Ni-K$\alpha$ to Fe-K$\alpha$ line intensity ratio are 0.066 and 0.026, respectively. Comparing this with the Monte-Carlo simulation of reflection, Ni to Fe abundance ratio is 1.9$\pm$0.8 solar. We discuss the results and possibility of Ni abundance enhancement., Comment: 16 pages, 4 figures, ApJ accepted

Objective: To evaluate the area of the midbrain and pons on mid-sagittal MRI in patients with progressive supranuclear palsy (PSP), Parkinson disease (PD), and multiple-system atrophy of the Parkinson type (MSA-P), compare these appearances and values with those of normal control subjects, and establish diagnostic MRI criteria for the diagnosis of PSP.Methods: The authors prospectively studied MRI of 21 patients with PSP, 23 patients with PD, 25 patients with MSA-P, and 31 age-matched normal control subjects. The areas of the midbrain tegmentum and the pons were measured on mid-sagittal MRI using the display tools of a workstation. The ratio of the area of the midbrain to the area of the pons was also evaluated in all subjects.Results: The average midbrain area of the patients with PSP (56.0 mm2) was significantly smaller than that of the patients with PD (103.0 mm2) and MSA-P (97.2 mm2) and that of the age-matched control group (117.7 mm2). The values of the area of the midbrain showed no overlap between patients with PSP and patients with PD or normal control subjects. However, patients with MSA-P showed some overlap of the values of individual areas with values from patients with PSP. The ratio of the area of the midbrain to the area of pons in the patients with PSP (0.124) was significantly smaller than that in those with PD (0.208) and MSA-P (0.266) and in normal control subjects (0.237). Use of the ratio allowed differentiation between the PSP group and the MSA-P group.Conclusion: The area of the midbrain on mid-sagittal MRI can differentiate PSP from PD, MSA-P, and normal aging.

The hard X-ray imager (HXI) and soft gamma-ray detector (SGD) onboard Astro-H observe astronomical objects with high sensitivity in the hard X-ray (5−80 keV) and soft gamma-ray (40−600 keV) energy bands. To achieve this high sensitivity, background rejection is essential, and these detectors are surrounded by large and thick bismuth germinate scintillators as an active shield. We have developed adequate trigger logic for both the HXI and SGD to process signals from main detector and BGO shield simultaneously and then we optimized the trigger delay and width, with consideration of the trigger latch efficiency. The shield detector system performs well, even after it is assembled as the HXI sensor. The energy threshold maintains the same level as that observed during the prototype development phase, and the experimental room background level of the main detector is successfully reduced by our optimized trigger timing.

The Soft Gamma-ray Detector (SGD) is a Si/CdTe Compton telescope surrounded by a thick BGO active shield and is scheduled to be onboard the ASTRO-H satellite when it is launched in 2015. The SGD covers the energy range from 40 to 600 keV with high sensitivity, which allows us to study nonthermal phenomena in the universe. The SGD uses a Compton camera with the narrow field-of-view (FOV) concept to reduce the non-Xray background (NXB) and improve the sensitivity. Since the SGD is essentially a nonimaging instrument, it also has to cope with the cosmic X-ray background (CXB) within the FOV. The SGD adopts passive shields called “fine collimators” (FCs) to restrict the FOV to ≤ 0.6° for low-energy photons (≤ 100 keV), which reduces contamination from CXB to less than what is expected due to NXB. Although the FC concept was already adopted by the Hard X-ray Detector onboard Suzaku, FCs for the SGD are about four times larger in size and are technically more difficult to operate. We developed FCs for the SGD and confirmed that the prototypes function as required by subjecting them to an X-ray test and environmental tests, such as vibration tests. We also developed an autocollimator system, which uses visible light to determine the transmittance and the optical axis, and calibrated it against data from the X-ray test. The acceptance tests of flight models started in December 2013: five out of six FCs were deemed acceptable, and one more unit is currently being produced. The activation properties were studied based on a proton-beam test and the results were used to estimate the in-orbit NXB.

Maximum likelihood (ML) estimates of K-distribution parameters are derived using the expectation maximization (EM) approach. This approach demonstrates the computational advantages compared with 2-D numerical maximization of the likelihood function using a Nelder-Mead approach. For large datasets, the EM approach yields more accurate estimates than those of a non-ML estimation technique.

This paper proposes a new formulation for speech recognition/understanding systems. In which the posteriori probability of a speaker's message that the speaker intends to address given an observed acoustic sequence is maximized. This is an extension of the current criterion that maximizes the probability of a word sequence. Among the various possible representations, we employ a co-occurrence score of words measured by mutual information as the conditional probability of a word sequence occurring in a given message. The word sequence hypotheses obtained by bigram and trigram language models are rescored using the co-occurrence score. Experimental results show that the word accuracy is improved by this method. Topic-words which represent the content of a speech signal are then extracted from speech recognition results based on the significance score of each word. When five topic-words are extracted for each broadcast-news article, 82.8% of them are correct in average. This paper also proposes a verbalization-dependent language model which is useful for Japanese dictation systems.

This paper proposes an efficient method for large-vocabulary continuous-speech recognition, using a compact data structure and an efficient search algorithm. We introduce a very compact data structure DAWG as a lexicon to reduce the search space. We also propose a search algorithm to obtain the N-best hypotheses using the DAWG structure. This search algorithm is composed of two phases: "forward search" and "traceback". Forward search, which basically uses the time-synchronous Viterbi algorithm, merges candidates and stores the information about them in DAWG structures to create phoneme graphs. Traceback traces the phoneme graphs to obtain the N-best hypotheses. An evaluation of this method's performance using a speech-recognition-based telephone-directory-assistance system having a 4000-word vocabulary confirmed that our strategy improves speech recognition in terms of time and recognition rate.

Formulated is a new systematic method for obtaining higher order corrections in numerical simulation of stochastic differential equations (SDEs), i.e., Langevin equations. Random walk step algorithms within a given order of finite $\Delta t$, are obtained so as to reproduce within that order a corresponding transition density of the Fokker-Planck equations, in the weak Taylor approximation scheme. A great advantage of our method is its straightforwardness such that direct perturbative calculations produce the algorithm as an end result, so that the procedure is tractable by computer. Examples in general form for curved space cases as well as flat space cases are given in some order of approximations. Simulations are performed for specific examples of U(1) system and SU(2) systems, respectively., Comment: 4 pages,latex,2 figures(GNU plot).Presented at LATTICE96(algorithms), the International Symposium on Lattice Field Theory, 4 - 8 June 1996, in St. Louis, Missouri, USA

Experimental data of\(\bar pp\) annihilation into 3π0 and π+π0π− are analyzed in the quark pair creation model (3P0 Model) with planar quark diagram topology. There are two form factors depending on whether 1) the two π′s contain a quark or an antiquark that originates from ap or a\(\bar p\) respectively, or 2) one π contain a quark and an anti-quark both created from the vacuum. We call the amplitude of the first as a G-type and the latter as the an F-type. The intermediate states in G-type processes which contain correlations of the3P0 vertices in the initial state and those in the final state can be regarded as gluons of 0++ and a\(\bar qq\). The two π resonance like structure of 0++ (1520) observed in 3π0 and 2++ observed in π+π0π− can be interpreted as the interference of the G-type amplitudes with the F-type amplitudes.

We construct an X-ray spectral model of reprocessing by a torus in an active galactic nucleus (AGN) with a Monte Carlo simulation framework MONACO. Two torus geometries of smooth and clumpy cases are considered and compared. In order to reproduce a Compton shoulder accurately, MONACO includes not only free electron scattering but also bound electron scattering. Raman and Reyleigh scattering are also treated, and scattering cross sections dependent on chemical states of hydrogen and helium are included. Doppler broadening by turbulence velocity can be implemented. Our model gives consistent results with other available models, such as MYTorus, except for differences due to different physical parameters and assumptions. We studied the dependence on torus parameters for Compton shoulder, and found that a intensity ratio of Compton shoulder to line core mainly depends on the column density, inclination angle, and metal abundance. For instance, an increase of metal abundance makes the Compton shoulder relatively weak. Also, shape of Compton shoulder depends on the column density. Furthermore, these dependences become different between smooth and clumpy cases. Then, we discuss the possibility of ASTRO-H SXS spectroscopy of Compton shoulder in AGN reflection spectra., Comment: 33 pages, 20 figures, ApJ in press

We report herein the case of a 43-year-old man in whom gastrointestinal bleeding was found to be caused by a large arteriovenous malformation (AVM) in the ileum. Emergency angiography proved invaluable in disclosing the AVM as the bleeding source, after which the patient was successfully treated by surgery. Although the number of reports of AVM has recently been increasing, ileal AVMs are still relatively rare. Moreover, according to the Japanese literature, no other patient has had such marked and large feeding arteries and draining veins as those in the patient described in this report.

We report three cases of cystic islet cell tumors, two caused by gastrinomas and the other by an islet cell carcinoma. All three patients underwent computed tomography (CT) and angiography and two also had magnetic resonance (MR) imaging. Several common radiographic findings were present as follows: thickening of the cyst wall and irregularity of the inner surface on postcontrast CT and MR images, neovascularity and a densely staining hypervascular rim on angiography, and moderately increased signal intensity of the cyst content on T1-weighted MR images, which indicated fluid containing blood and/or necrotic tissue. These findings are helpful in defining the nature of these lesions and strongly suggest that cystic islet cell tumor should be included in the differential diagnosis.

We have developed an embolizing material consisting of Y-shaped silicone particles for partial splenic embolization. Wide spaces for blood flow are left around the particles when these are lodged in arterial branches. We embolized one kidney in each of 3 dogs with the particles and observed a slowly induced occlusion of renal arterial branches during one month. The particles were also used for partial splenic embolization in 14 patients. The average portion of infarcted spleen tissue 7 days after embolization was 51% calculated from CT images. In 6 patients who had CT both 2 and 7 days after embolization, the average rate of splenic infarction increased from 29% at 2 days to 60% at 7 days. Our patients required analgesics for only 2.3 days on average. The Y-shaped silicone particles by slowly occluding splenic arterial branches produce ischemia in a gradual fashion which may minimize the pain after embolization.

IVL is characterized by a propensity for intravascular tumor cell proliferation. Premortem diagnosis of IVL is difficult because of its nonspecific clinical, laboratory, and imaging manifestations. This study examined cerebral MR imaging patterns of IVL and their changes with and without chemotherapy. Nine of 11 patients studied presented with abnormal findings. We define 5 patterns of abnormal MR imaging findings: 1) infarctlike lesions, 2) nonspecific white matter lesions, 3) meningeal enhancement, 4) masslike lesions, and 5) hyperintense lesions in the pons on T2WI. Seven patients presented with only 1 pattern, while 2 patients presented with multiple patterns. Lesions in 7 treated patients responded to chemotherapy. Pathologic specimens revealed intravascular tumor cell infiltration with associated infarctions, necrosis, congestion, demyelination, vasculitis, and tumor cell extravasation. We conclude that MR imaging patterns can be possible manifestations of intravascular-dominant infiltration by tumor cells with associated occlusion or inflammation, depending on the level of affected vessels.

The MRI appearances of calcification within intracranial tumours was assessed by reviewing MR images of 11 calcified tumours documented by CT. The signal intensity of the calcified regions was varied and nonspecific on both T1- and T2-weighted images. They were seen as signal void exclusively on T2-weighted images in only 2 patients. Gadolinium enhancement of the calcified portion occurred in 7 of 10 patients. These findings reflect the presence of tumour parenchyma within the calcified region, as proved in 5 lesions examined histologically.

Correlation of magnetic resonance (MR) imaging findings and those at pathologic evaluation was attempted in six cases of solid and papillary epithelial neoplasm of the pancreas. All patients were female, and the mean patient age was 26 years (range, 13-73 years). On T1-weighted spin-echo images, tumors were well demarcated, and areas of high signal intensity were evident within them. At macroscopic examination, these areas corresponded to solid portions with marked hemorrhagic necrosis or cystic portions filled with hemorrhagic debris. In three of four masses surrounded by macroscopically evident fibrous capsules, a rim of low intensity was revealed at T1-weighted imaging. When T1-weighted spin-echo imaging reveals obvious areas of high intensity within a sharply marginated tumor of the pancreas, especially in a young woman, solid and papillary epithelial neoplasm might be a primary diagnostic consideration.

Dieulafoy type ulcer was originally reported as a kind of characteristic ulcers of stomach which can cause massive bleeding. A recent experience with massive jejunal bleeding from tiny ulcer of Dieulafoy type as well as some device identifying the bleeding site intraoperatively is described.A 39-year-old woman was admitted because of abdominal pain and large amount of melena. Angiography revealed massive bleeding from a peripheral region of the first branch of the jejunal temporarily artery into jejunal lumen. To control the massive bleeding, transarterial embolization (TAE) using metallic coil was performed. Since precise bleeding point was accurately identified during operation by means of radiographic device using metallic coil for TAE as a landmark, only small resection of the jejunal wall was performed. Pathological examination revealed characteristic features of Dieulafoy type ulcer such as small shallow ulcer, rupture of persistent caliber artery, and middle-sized serpentine arteries in the submucosal layer.

The author reviews major methods of applying the vector quantization (VQ) technique to speech and speaker recognition. These include speech recognition based on the combination of VQ and the DTW/HMM (dynamic time warping/hidden Markov model) technique. VQ-distortion-based recognition, learning VQ algorithms, speaker adaptation by VQ-codebook mapping, and VQ-distortion-based speaker recognition. It is concluded that not only has the VQ technique reduced the amount of computation and storage, but it has also created new ideas for solving various problems in speech/speaker recognition. >

The authors modified Gianturco expandable metallic stents to make them relocatable. Two tandem designs were made. The first design had long struts and, thus, more evenly distributed expansive force; the second had short struts and more flexibility and was better suited for use in curved strictures. Both designs (two to four in tandem, 0.012-, 0.014-, and 0.018-inch wire) were tested successfully in vitro, and both designs (two to four in tandem, 0.012-inch wire) were successfully placed, retrieved, and relocated in the inferior vena cava of five dogs.

Infantile hemangioendothelioma of the liver is rare disease, and only 31 cases have been reported in the last 7 years in Japan. A case of giant hemangioendothelioma of the liver in a 2-month-old female is reported, in which differential diagnosis from blastoma of the liver was difficult by imaging methods. The patient was pointed out to have an enlarged liver at the first month examination, and was admitted to examine for the enlarged liver.On admission, the liver was enlarged and the liver edge was palpated at 3 fingerbreadth below the costal margin. Right lobectomy was performed under a tentative diagnosis of hepatoblastoma resulting from US, CT, and angiography. The final pathological diangosis revealed it to be hemangioendothelioma. In this case preoperative definite diangosis of hemangioendothelioma was difficult because no symptoms such as congestive heart failure and skin lesion were associated. There exist two therapeutic methods for the disease, namely conservative and surgical treatments. Surgical intervention may be recommended for a case in which the tumor forms a large mass with a potential malignancy like this case.

Within the framework of the constituent quark model we discuss the effects of different types of meson-baryon-baryon vertex operators on the form factors and the coupling strengths of the lowest-lying positive and negative parity non-strange baryons. We compare the quark pair creation model (3P0-model) with the SU(6)-model in which mesons are treated as elementary fields that directly couple to the quarks. The latter model is employed both in the so-called static limit and in a modification motivated by Galilei invariance. It is demonstrated that the inclusion of non-static effects simulates some features of the3P0 vertex. Especially the reaction πN→ππN is found to be very sensitive to the different assumptions on the dynamics of theq¯q pair creation process. More indirect hints for the internal structure of the mesons might be obtained from the predicted asymmetry for the two form factors ofN→Δ+π andΔ→N+π, which occurs in the3P0-model, only.

\(N\bar N\) annihilation into three pseudoscalar mesons especially πππ and\(K\bar K\pi \) are studied in the quark pair creation model or the3P0 model in which two\(q\bar q\) pairs are annihilated and two\(q\bar q\) pairs are created with quantum numbers of the vacuum or3P0. The correlations of two pions to form ϱ,f2 and the resonance AX(1565) which is recently found by the ASTERIX group are taken into account. A proper treatment of the symmetry among the three pions in the final state shows that the\({}^{31}S_0 p\bar p\) annihilation into ϱπ is suppressed in agreement with the experiment. We calculate the cosθ distribution or the distribution of the Dalitz plot as the function of the angle between the direction of emission of one decay pion in the resonance centre of mass and the line of flight of the resonance. The interferences of π+ρ+, π−ρ+ and π0ρ0 in the isospin 0 channels and π±ρ∓ and π0f2 in the isospin 1 channel reproduce the peaks. The cos θ distribution for the P-wave\(p\bar p\) annihilation into πAX depends strongly on the size of the pion since the amplitude interfers with the π±ρ∓ amplitude which is sensitive to the size of the pion. The same model qualitatively explains the\(p\bar p\) annihilation into\(K\bar K\pi \) in whichK or\(\bar K\) and π are correlated to formK*\(\bar K\) or\(\bar K\)*K final states. We can qualitatively reproduce different patterns of the cos θ distribution for theK*+ andK*0.

A vector quantization (VQ)-based recognition method which uses feature vector codebooks containing hierarchical spectral dynamics is proposed. This method is highly effective for reducing the number of candidates in word recognition and achieving a high recognition accuracy in /b/,/d/ and /g/ recognition. Since this method does not need time alignment, it has the advantage of a small amount of computation and ease of parallel processing. Experimental results comparing the performances of the multiple-codebook and single-codebook methods indicate that, when the codebook size is small, the multiple-codebook method is better than the single-codebook method. However, if the codebook size is reasonably large, the single-codebook method displays better performance than the multiple-codebook method. >

The International Speech Communication Association (ISCA) is a worldwide nonprofit organization that supports research in all aspects of spoken-language communication. The Association was founded in 1988 as the European Speech Communication Association (ESCA). It is incorporated in France, governed by a 13-member board, composed of members from all over the world.

Speaker recognition is the process of recognizing the speaker using speech signals, which can be classified into speaker identification and verification. Speaker identification is the process of determining from which of the registered speakers a given utterance comes. Speaker verification is the process of accepting or rejecting the identity claimed by a speaker. Speaker recognition can also be classified into text-dependent, text-independent, and text-prompted methods. Spectral envelope and prosody features of speech are usually used as speaker features. In order to accommodate intraspeaker variations of signal characteristics, it is important to apply parameter-domain and/or likelihood-domain normalization/adaptation techniques. High-level features, such as word idiolect, pronunciation, phone usage, and prosody, have recently been investigated in text-independent speaker verification.

New techniques for automatic speaker verification are described which are based on cepstral analysis of fixed sentence-long utterances. Utterances are represented by time functions of cepstral coefficients expanded by an orthogonal polynomial. Sample utterance representations are compared with similarly constituted reference patterns to decide whether to accept or reject an identity claim. A dynamic programming technique is used to bring sample utterance representations into time registration with reference patterns. A local distance measurement is summed along the optimal path associated with the time registration to provide an overall distance which is compared with a threshold to effect the decision. Results of an evaluation using prerecorded utterance sets including telephone speech are presented. Verification error rates less 1% are obtained.

This paper proposes the application of a tree-structured clustering method that integrates the effects of noise as well as SNR variation in the framework of piecewise-linear transformation (PLT)-based noise adaptation for robust speech recognition. According to the clustering results, a noisy speech HMM is made for each node of the tree structure. An HMM that best matches the input speech is selected based on the likelihood maximization criterion by tracing the tree downward from the top (root), and the selected HMM is further adapted by linear transformation. The proposed method is evaluated by applying it to a Japanese dialogue recognition system. Experimental results confirm that the proposed method is effective in recognizing numerically noise-added speech and actual noisy speech uttered by a wide range of speakers under various noise conditions.

Dissection of the superior mesenteric artery (SMA) not associated with aortic dissection is rare. The purpose of this study is to describe the computed tomographic (CT) findings of this condition. We studied the CT findings of six patients with isolated dissection of the SMA. CT demonstrated thrombosis of the false lumen or intramural hematoma (n = 4) and/or intimal flap (n = 4) in all six patients. Other CT findings were enlarged diameter of the SMA (n = 5), increased attenuation of the fat around the SMA (n = 5), and hematoma in the mesentery with hemorrhagic ascites (n = 1). CT is useful for the diagnosis of isolated dissection of the SMA, and increased attenuation of the fat around the artery is considered the key to the diagnosis when no definite findings are evident.

How to recognize and understand spontaneous speech is one of the most important issues in state-of-the-art speech recognition technology. In this context, a five-year large scale national project entitled "Spontaneous speech: corpus and processing technology" started in Japan in 1999. This paper gives an overview of the project and reports on the major results of experiments that have been conducted so far at Tokyo Institute of Technology, including spontaneous presentation speech recognition, automatic speech summarization, and message-driven speech recognition. The paper also discusses the most important research problems to be solved in order to achieve ultimate spontaneous speech recognition systems.

This contribution -- a continuation of earlier work -- reports on recent developments in the calculation and understanding of 4-quark energies generated using lattice Monte Carlo techniques., Comment: 3 pages, latex, no figures, contribution to Lattice 94

The authors propose a new mathematical method for extracting spectral movement in a time-sequence of speech spectrum. The spectral movement is characterized by the time and frequency derivative of a time-sequence of log spectrum envelopes. Spectral movement direction, the movement toward a higher or lower frequency region, can be identified by the sign of the proposed function. A parameter which can be used for speech segmentation is derived form this function A distance measure for speech recognition is also derived as the Euclidean distance between two spectral movement patterns extracted by the proposed function. This distance is easily calculated using cepstrum coefficients. Speech recognition results using dynamic time warping template matching with this new distance measure indicate that recognition error rate can be reduced to less than half compared with the conventional Euclidean cepstrum distance measure. >

The paper describes a way of improving extended HMM composition that can precisely adapt HMMs to both noisy and distorted speech. To do this, the authors incorporate the variance of power into extended HMM composition using quantization to approximate the Gaussian distribution of the 0th order cepstrum. Consequently, a distribution of noisy speech is approximated in the linear spectral domain as a mixture of log normal distributions. This method is evaluated by a four-digit recognition experiment when the number of digits is known. Two types of noise, computer room noise and car noise, are used and noisy and distorted speech data is made by adding these types of noise to speech data recorded using a boundary microphone. Results show that the proposed method improves recognition rates for noisy and distorted speech compared with their previous method.

Studies Japanese large-vocabulary continuous-speech recognition (LV CSR) for a Japanese business newspaper. To enable word N-grams to be used, sentences were first segmented into words (morphemes) using a morphological analyzer. About five years of newspaper articles were used to train N-gram language models. To evaluate our recognition system, we recorded speech data for sentences from another set of articles. Using the speech corpus, LV CSR experiments were conducted. For a 7k vocabulary, the word error rate was 82.8% when no grammar and context-independent acoustic models were used. This improved to 20.0% when both bigram language models and context-dependent acoustic models were used.

The paper describes major research and development in speech recognition and synthesis technologies at NTT from the telecommunications applications viewpoint. Technologies include speaker-dependent, speaker-independent word recognition based on DP matching, speaker-independent word spotting based on HMM, large vocabulary speaker-independent continuous speech recognition based on HMM-LR and high-quality Japanese text-to-speech synthesis. A commercial ANSER system that uses speech recognition and synthesis technologies is also introduced. >

This paper overviews the speech recognition issues for the Japanese language, and introduces three research projects conducted at NTT (Nippon Telegraph and Telephone) Human Interface Laboratories and ATR Interpreting Telephony Research Laboratories. The first topic is stochastic language models for the sequences of Japanese characters to be used in a Japanese dictation system with unlimited vocabulary. The second topic is an accurate and efficient algorithm for very-large-vocabulary continuous speech recognition based on an HMM-LR algorithm. This algorithm was applied to a telephone directory assistance system that recognizes spontaneous speech having a vocabulary size of roughly 80000. The third topic is a continuous speech recognition system based on strategies of the phoneme-context-dependent (allophonic) modeling and parsing. >

A 67-year-old man was hospitalized at our institution complaining with epiphora and exophthalmos on the left side. Magnetic resonance imaging (MRI) showed an ill demarcated retrobulbar mass in the left orbit. 67Ga scintigraphy revealed avid uptake in the left orbital region. The patient was treated with radiation therapy. One month after the radiation therapy, the size of the mass decreased remarkably on MRI. 67Ga planar imaging after treatment showed no uptake, but 67Ga SPECT showed slightly increased uptake in the left orbital region. One year after the radiation therapy, MRI showed residual mass in the left orbital region. Both 67Ga planar imaging and SPECT showed no uptake in the left orbital region. 1.8 years after the radiation therapy, MRI showed the residual mass with no interval change in size. Both 67Ga planar imaging and SPECT showed no uptake in the left orbital region. The patient remains well with no evidence of local recurrence. 67Ga scintigraphy is useful in assessing the response to radiation therapy of MALT lymphoma in this case.

To evaluate radiation exposure to patients and radiologists during transcatheter arterial embolization(TAE) for hepatocellular carcinoma.In 39 TAE procedures performed at eight institutes, skin doses were evaluated with thermoluminescence dosimeters at the patient's back(entrance surface) and lower abdomen, and at the radiologist's forehead and abdomen. Real-time dosimeters were also used to evaluate patient skin dose.The patients' mean entrance surface dose was 973 +/- 681 mGy(range, 185 to 3543 mGy) with the mean fluoroscopic time of 21 minutes and 6 digital subtraction angiography(DSA) acquisitions. The dose at the patients' lower abdomen was 0.98 +/- 0.77 mGy. Doses for the radiologists were 0.04 +/- 0.04 mGy at the forehead and 0.15 +/- 0.19 mGy and 0.005 +/- 0.01 mGy at the abdomen over and under the apron, respectively. Fifty-six percent of the patients' skin dose was from DSA and 44% from fluoroscopy.Patient skin dose may occasionally exceed the dose for transient erythema. Because a patient may have repeated TAEs, skin doses or X-ray conditions should be recorded. The exposed doses of radiologists were considered to be acceptable with proper techniques. Further efforts to reduce radiation should be directed toward both DSA and fluoroscopy.

Wire-directed detachable balloons were used in embolization treatment of varicoceles in six patients. Advancement of the balloon over a guide wire allowed easy embolization of the internal spermatic vein at the level of the inguinal ring. Additional detachable balloons were used to occlude the collateral vessels to the varicocele. This method proved highly effective in embolization of tortuous veins and in placement of multiple balloons.

Preoperative transhepatic portal vein embolization may not always be sufficient to achieve the desired changes in contralateral hepatic volume and function. The beneficial role of additional transcatheter arterial embolization performed after inadequate response to preoperative transhepatic portal vein embolization is described.Four patients underwent both preoperative transhepatic portal vein embolization and transcatheter arterial embolization, and 6 control patients underwent preoperative transhepatic portal vein embolization only. Changes in right liver lobe volume fraction, residual left lobe volume fraction, and prediction score (low-risk,45; borderline, 45-55; high-risk55); were evaluated.1) The change in right liver lobe volume after both preoperative transhepatic portal vein embolization and transcatheter arterial embolization (volume after/before) was 0.75 times that of the original level whereas after preoperative transhepatic portal vein embolization, they were only 0.81 times that of the original level. 2) The change in residual left liver volume after both preoperative transhepatic portal vein embolization and transcatheter arterial embolization (volume after/before) was 1.40 times that of the original level whereas after preoperative transhepatic portal vein embolization they were only 1.30 times than the original level. The changes in left liver volume after preoperative transhepatic portal vein embolization/transcatheter arterial embolization was more favorable than those after preoperative transhepatic portal vein embolization only. 3) The change in prediction score after both preoperative transhepatic portal vein embolization and transcatheter arterial embolization (after/before) was 0.81 times that of the original level. All prediction score in high-risk patients recovered to the borderline or safety zone. Change after preoperative transhepatic portal vein embolization only (before/after) was 0.87 times that of the original level. 4) All 4 patients who underwent both preoperative transhepatic portal vein embolization and transcatheter arterial embolization received right hepatic lobectomy successfully and returned to their normal life style.Preoperative occlusion of right hepatic inflow vessels increased the volume and function of the contralateral lobe where high-risk patients recovered to the borderline zone for major hepatic resection.

Contrast enhanced CT was performed in seven patients with pancreatic disease (chronic pancreatitis, n = 3; pancreatic head cancer, n = 2; mucin-producing pancreatic tumor, n = 2) who showed dilation of the main pancreatic duct (MPD) Minimum intensity projection (Min-IP) images of the pancreas were obtained using multi-projection volume reconstruction (MPVR) software by selecting an oblique slab that contained the entire MPD. Curved reformation (CR) images were obtained using multiplanar reformation (MPR) software by tracing the MPD on the Min-IP image. Both Min-IP images and CR images clearly showed the dilated main pancreatic duct in all seven patients. In three of the seven, obstruction of the MPD in the pancreatic head and the cause of obstruction (tumor mass, n = 2; caliculus, n = 1) were also clearly seen. Min-IP and CR images seem to be useful for the diagnosis of pancreatic diseases.

1H-NMR spectroscopy was performed on a 6-month-old boy with pachygyria. A MRI study demonstrated an abnormally thick cortex localized in the right occipital lobe. 1H-NMR spectrums were collected from the lesion and the contralateral cortex that appeared normal on MRI. The N-acetylaspartate (NAA)/Cre (creatine) ratio was markedly lower in the abnormal cortex than in the contralateral cortex and the occipital cortex. NAA localizes to neurons, axons, dendrites and synaptic connections and increases with maturation of neurons. Its decrease is considered to represent the decrease in the number of these structures and/or disturbance of neuronal maturation. We conclude that NAA/Cre can be an important index that reflects the pathogenesis of pachygyria.

The purpose of this study was to evaluate the brain perfusion of neuro-Behçet disease using 99mTc-HMPAO. In total, nine patients (7 men and 2 women) were evaluated. Seven were in the chronic phase, one was acute and the other was subacute. The results of perfusion study were compared with the MRI findings. The findings of chronic-phase neuro-Behçet disease were as follows: 99mTc-HMPAO showed hypoperfusion of the cerebellum in 3, brain stem in 5, and cerebral cortex in 4 patients. Moreover, asymmetry of the basal ganglia and thalamus was seen in 4 cases with 99mTc-HMPAO. MRI showed atrophy of the cerebellum in 4, brain stem in 5 and cerebral cortex in 2 patients. One acute patient showed a perfusion defect with 99mTc-HMPAO that was consistent with the vasculitis shown with MRI. In the remaining subacute patient, perfusion defect was demonstrated with 99mTc-HMPAO, consistent with the infarction shown with MRI. It is concluded that brain perfusion scintigraphy is useful to evaluate the brain blood flow of neuro-Behçet disease.

A cDNA encoding spinach alpha-glucosidase was cloned and sequenced by the reverse-transcription polymerase chain reaction (RT-PCR) and rapid amplification of cDNA ends (RACE) methods. The cDNA comprised 2867 bp, and included an open reading frame which encodes a polypeptide of 903 amino acid residues. The calculated molecular mass of 101 kDa was larger than those of native alpha-glucosidases in spinach seeds, which are 78, 78, 82, and 82 kDa by SDS-PAGE for alpha-glucosidase I, II, III, and IV, respectively. The deduced amino acid sequence included those of tryptic peptides from native enzymes. Southern blot analysis suggested that the alpha-glucosidase gene was a single-copy gene. These results indicate the possibility that the multiplicity of alpha-glucosidase in spinach occurs via post-translational modification.

For the evaluation of brain tumor (n = 15), we performed both dynamic and static 99mTc-ECD (ECD) SPECT studies 201Tl SPECT was also used for comparison with the results of ECD SPECT. Dynamic ECD SPECT was obtained following the injection of 600 MBq of ECD. Five min after the injection of ECD, static ECD SPECT was performed. 201Tl SPECT was obtained 10 min after the injection of 74MBq. Abnormal uptake was recognized in 7 of 15 tumors with dynamic ECD: 5 of 7 meningiomas, 1 of 1 glioblastoma and 1 of 1 astrocytoma. However, no abnormal uptake was seen in 3 of 3 benign tumors (1 low grade astrocytoma, 1 hemangioma, 1 cranio-pharyngioma) and in 2 of 2 brain metastases. In contrast abnormal uptake was seen in 11 of 15 tumors with 201Tl:7 of 7 meningiomas; 2 of 2 brain metastases, 1 of 1 glioblastoma and 1 of 1 craniopharyngioma. No abnormal uptake was seen in 3 of 3 benign tumors (1 hemangioma and 2 low grade astrocytomas). Equivocal uptake was seen in 1 low grade astrocytoma with dynamic ECD and 201Tl. The mechanism of the accumulation of dynamic ECD to brain tumor is unclear. However, it may reflect not only blood flow, but also metabolism.

Helical computed tomography was performed with intravenous contrast material administration in ten patients with hepatic tumors (metastatic tumor, n = 6; hepatocellular carcinoma, n = 4). Maximum intensity projection (MIP) images of intrahepatic portal venous branches and hepatic veins, and shaded surface display (SSD) images of the hepatic tumors were reconstructed for each patient. The MIP image and SSD image were fitted to overlap on a single display to obtain three-dimensional (3-D) images showing both the hepatic vessels and hepatic tumors. Subsegmental localization of the tumors determined from the 3-D images proved to be correct in all of the six patients who received hepatic resection after examination. 3-D images seem to be useful for preoperative assessment of hepatic tumors.

The existence of diabetic cardiomyopathy has been substantiated in patients with diabetes mellitus. In this study we evaluated two mice with diabetes mellites (DM) and two control mice to determine whether the fatty acid metabolism of the myocardium is abnormal when 125I-BMIPP is used. 125I-BMIPP of 0.74 MBq (20 microCi) was injected, and 30 min later, the mice were slaughtered. Heart, lungs, liver, spleen and kidneys were removed. After freezing the heart, myocardium was sliced into 20 micron sections using a cryostat. Each slice was placed in contact with an imaging plate for autoradiography. A bioimaging analyzer was used to analyze the data. The percent injected dose/g of heart was lower in the DM group than in the control group (11.6 +/- 0.49 vs 18.3 +/- 1.27). Each slice of the DM group also had fewer counts than slices in the control group (p0.05 or p0.01). These results indicated that fatty acid metabolism in the myocardium was less in the DM group than in the control group.