森本 雅和

【目的】本研究では,2型糖尿病が強く疑われる人とそうでない人の生活習慣の比較を行うことを目的とした。【方法】A市において2008年度から2019年度までに国民健康保険の健康診査を受診した30歳以上の男女のうち,解析項目に欠損がある人,既に生活習慣改善に取り組んでいる人を除く10,401名を解析対象とした。糖尿病が強く疑われる人を「HbA1c(NGSP値)が6.5%以上かつ/または血糖値を下げる薬の服薬ありの人」と定義し,χ2検定を用いて糖尿病の疑いの有無の2群間で生活習慣の比較を行った。その後,糖尿病が強く疑われる/それ以外を従属変数,各生活習慣項目を独立変数,性別と年齢で調整した二項ロジスティック回帰分析を行った。解析は65歳未満と以上で行った。【結果】二項ロジスティック回帰分析の結果,65歳未満の対象者においては,日常生活において歩行または同等の身体活動を1日1時間以上実施していない人,他の人と比較して食べる速度が速い人,就寝前の2時間以内に夕食をとることが週に3回以上ある人は,そうでない人と比較して糖尿病が強く疑われる人のオッズ比がそれぞれ1.69倍(95%信頼区間:1.22-2.32),1.31倍(1.01-1.72),1.52倍(1.12-2.07)であった。一方,お酒を時々あるいは毎日飲み,かつ1日当たり1～2合飲む人では,ほとんど飲まない・飲まない人と比較して糖尿病が強く疑われる人のオッズ比が0.65倍(0.46-0.92)であった。65歳以上の対象者では,日常生活において歩行または同等の身体活動を1日1時間以上実施していない人,ほぼ同じ年齢の同性と比較して歩く速度が速くない人,他の人と比較して食べる速度が速い人は,そうでない人と比較して糖尿病が強く疑われる人のオッズ比がそれぞれ1.41倍(1.18-1.70),1.21倍(1.01-1.45),1.29倍(1.09-1.53)であった。【結論】本研究より,糖尿病予防のためには65歳未満,65歳以上の人に共通して身体活動を行うこと,早食いをしないことが重要であり,加えて65歳未満の人では就寝前2時間以内の夕食摂取に注意すべきことが示唆された。(著者抄録)

Correct assessment of the bone healing process is required for the management of limb immobilization during the treatment of bone injuries, including fractures and defects. Although the monitoring of bone healing using ultrasound poses several advantages regarding cost and ionizing radiation exposure compared with other dominant imaging methods, such as radiography and computed tomography (CT), traditional ultrasound B-mode imaging lacks reliability and objectivity. However, the body structures can be quantitatively observed by ultrasound frequency-based methods, and therefore, the disadvantages of B-mode imaging can be overcome. In this study, we created a femoral bone hole model of a rat and observed the bone healing process using the quantitative ultrasound method and micro-CT, which provides a reliable assessment of the tissue microstructure of the bone. This study analyzed the correlation between these two assessments. The results revealed that the quantitative ultrasound measurements correlated with the CT measurements for rat bone healing. This ultrasound frequency-based method could have the potential to serve as a novel modality for quantitative monitoring of bone healing with the advantages of being less invasive and easily accessible. Impact statement Bone healing monitoring with ultrasound is advantageous as it is less invasive and easily accessible; however, the traditional B-mode method lacks reliability and objectivity. This study demonstrated that the proposed ultrasound frequency-based monitoring method can quantitatively observe bone healing and strongly correlates with the computed tomography measurements for rat bone healing. This method has the potential to become a reliable modality for monitoring bone healing.

© 2018 IEEE. In this paper, we propose a new approach to classify ovarian follicles into two classes. A smoothing filter which is designed to consider speckle patterns under the resolution of the ultrasound devices is applied for filtering ovarian follicle images. Then, convolutional neural networks are used for extracting features from the filtered ovarian follicle images. Finally, both features extracted by CNNs from the filtered ovarian follicle images and numerical features defined by our previous works are used for classification. From experimental results, we show the effectiveness of our proposed method.

© 2018 IEEE. In order to estimate the existence of ovum in ovarian follicle from ultrasound images, we reconstruct three-dimensional follicle model by using acceleration sensor. Because the ovum is too small to see in ultrasound image, so we are going to estimate its existence from shape and texture features of the follicle by using some AI methods. First, we obtain ultrasound images and rotation angle of probe simultaneously by attaching acceleration sensor on the probe. Then, we reconstruct the follicle as three-dimensional point cloud information. From the follicle point clouds, we can achieve arbitrary cross-sectional images. Then, we extracted shape feature quantity and compare them with actual cross-sectional image. As a result, we showed that we can extract shape feature quantities in arbitrary cross section. We also clarified that if the rotation speed of the ultrasonic probe is sufficiently slow, it can reproduce texture features as well as shape features.

Navigation applications for smartphones are poplar system, recently. Especially, WLAN Finger Print technique is suitable for indoor environment where GPS is difficult to use. This contribution describes a Finger Print Localization scheme using Deep Learning technique. First, the principle and experimental results of Finger Print using Deep Learning are described. Second, Coarse-to-Fine Localization based on SOM is proposed. A scheme to guess ahead accuracy for WLAN/GPS switching is described in the last section.

Copyright © 2017 The Institute of Electronics, Information and Communication Engineers. To promote effective guidance by health checkup results, this paper predict a likelihood of developing lifestyle-related diseases from health check data. In this paper, we focus on the fluctuation of hemoglobin A1c (HbA1c) value, which deeply connected with diabetes onset. Here we predict incensement of HbA1c value and examine which kind of health checkup item has important role for HbA1c fluctuation. Our experimental results show that, when we classify the subjects according to their gender and triglyceride (TG) fluctuation value, we will effectively evaluate the risk of diabetes onset for each class.

To promote effective guidance by health checkup results, this paper predict a likelihood of developing lifestyle-related diseases from health check data. In this paper, we focus on the fluctuation of hemoglobin A1c (HbA1c) value, which deeply connected with diabetes onset. Here we predict incensement of HbA1c value and examine which kind of health checkup item has important role for HbA1c fluctuation. Our experimental results show that, when we classify the subjects according to their gender and triglyceride (TG) fluctuation value, we will effectively evaluate the risk of diabetes onset for each class.

To realize a practical bread recognition system, we introduce RGB-D sensor to recognize partially overlapped breads. In this paper, we focus on upper objects, which leans on other objects. Appearance of the object changes significantly, so it degrades recognition accuracy. In the proposed method, we adjust the superior object's appearance by using depth sensor, to improve recognition accuracy.

<p>Total knee Arthroscopy (TKA) is an operation which replaces the damaged knee with an artificial knee implant. There are some kinds of TKA procedures, and various kinds of prosthesis. Thus, it becomes a tough work for surgeons to select an appropriate procedure and prosthesis for individual patients. This study proposes a prediction method of post-operative implanted knee kinematics. It predicts the post-operative kinematics from only pre-operative kinematics using a machine learning method with clinical big data. In 46 TKA subjects, the method predicts the post-operative anterior-posterior translation with a correlation coefficient of 0.77 and a root-mean-squared error of 0.7mm.</p>

To realize automated checker system of cafeteria, we have developed a meal menu recognition system. This system employs a RGB-D sensor to get both color image and depth image. From these images, first we extract and distinguish each dishes. By knowing what type of dish is used for the meal, we can reduce candidates for meal menus, it leads up to improvement of menu recognition accuracy. Some experimental result show that we can recognize dish type over 99% accuracy except for black dishes. Further, we can recognize a dish menu about 97% accuracy.

It is known that lifestyle habit and genetic factor are main reasons that occur cerebral aneurysms. In addition, some studies suggest that cerebral artery shape might be correlated with a risk of occurring aneurysms. For the purpose of preemptive medical care of the cerebral aneurysm, this study proposes a method to estimate a risk of occurring cerebral aneurysms based on the cerebral artery structure. The method extracts morphometric features of the Wills ring such as 3-D artery shape and bifurcation angle in 3-D magnetic resonance angiography (MRA) images. It then estimates the risk of occurring cerebral aneurysms from the extracted features using support vector machines (SVM). To validate the proposed method, we employed 40 subjects with cerebral aneurysms, and 40 subjects without cerebral aneurysms. Leave-one-out cross validation test was performed, and the method using 3-D artery shape achieved a sensitivity of 75% and a specificity of 75%; one using bifurcation angle did a sensitivity of 33% and a specificity of 71%; one using all features did a sensitivity of 68% and a specificity of 89%. The results showed that 3-D shape is effective for cerebral aneurysm occurrence risk prediction.

In the infertility treatment, medical examinations using ultrasonic devices, which can diagnose the mother's body safely on real time, are major. It is very difficult to judge the existence of an ovum before carrying out the paracentesis. A system which can distinguish the existence of the ovum in the ovarian follicle using ultrasonic devices is required. In this paper, several features of the deformation of ovarian follicle are defined and extracted from the ultrasonic moving image in a paracentesis operation. These features are extracted from the ultrasonic moving image obtained at the time of an ovum extraction operation. We investigate whether some clusters according to the existence of the ovum are formed using the defined features. Moreover, we also investigate whether some tendency exists in these features by the existence of an ovum.

Anterior cruciate ligament (ACL) injury causes knee joint instability, and affects on sports performance. Therefore, ACL reconstruction is essential to keep their performance high and to prevent osteoarthrosis. It is well known that the outcome of ACL reconstruction is strongly related to the placement and orientation of the bone tunnel. 2-D X-ray radiograph and CT images have been used to evaluate the placement and orientation of the bone tunnel. Quadrant method evaluates the bone tunnel placement based on the Blumensaat's line which has high intensity on 2-D X-ray lateral radiograph. There is problem of invasiveness using X-ray radiograph or CT image. Therefore, we have proposed an MR image based computer-aided surgical planning of ACL reconstruction. The system evaluates the bone tunnel placement and orientation based on Quadrant method. The remained problem of our system is Blumensaat's line is manually determined. This paper proposes that a method to synthesize the pseudo lateral radiograph from MR images, and extract the Blumensaat's line on the synthesized pseudo lateral radiograph. The experimental results showed that the proposed method successfully determined the Blumensaat's line on the pseudo lateral radiograph.

To improve object recognition accuracy, we introduce RGB-D sensor to get depth features and we construct two-stage recognition which combine recognition method with different characteristics. Experimental result show that the combination of these two refinement is effective to distinguish similar objects.

To realize object recognition in practical applications, it is necessary to handle layered objects. In this paper we introduce RGB-D sensor for object recognition system to treat overlapped objects. In the proposed method, we divide objects into segments and merge them by consulting partially recognition scheme. By several experiments, out method can recognize 30% occluded object with enough accuracy.

前十字靭帯(ACL;Anterior cruciate ligament)の損傷により出現した膝関節の不安定性はスポーツ成績に悪影響を与えるため，ACLを再建し安定した膝を作ることが必要である．ACL再建術の術後成績はACL挿入のための骨孔の位置と角度に強い相関があることが知られているため，骨孔開口の最適化は術後成績向上の重要な因子である．しかし，現在の術式では医師が内視鏡下で骨孔位置と角度を決定しているため，血管や靭帯を傷つける可能性があり，これがACL再建術の正確性や再現性を低下させていると考えられる．ACL再建術の術後評価法として単純X線画像を用いて骨孔位置，角度を評価するQuadrant法がある．本研究では，MR画像を用いた骨孔の位置と角度の評価を行う計算機診断支援システムを提案する．提案手法ではMR画像を用いたQuadrant法により，大腿顆周辺の血管や靭帯の3次元位置を考慮した骨孔位置を決定する．本システムをACL損傷患者に適用した結果，正確性と再現性の向上が確認できた．

The neonatal cerebral disorders severly languish the quality of life (QOL) of patients and also their families. It is required to detect and cure in their early stage for the sake of decreasing the degree of symptoms. However, it is difficult to evaluate neonatal brain disorders based on morphological analysis because the neonatal brain grows quickly and the brain development progress is different from person to person. Previously, we proposed a method of calculating growth index using Manifold learning. The growth index is effective to evaluate the brain morphological development progress, although, it does not directly correspond to the brain development delay. To evaluate brain development delay, this paper proposes an estimation method of neonatal brain age using Manifold learning, principal component analysis, and multiple regression model. The regression model is trained using a 4-D standard brain, which is constructed using training subjects with growth index. To evaluate the proposed method, we constructed a multiple regression model using 11 normal subjects (revised age: 0-4 month old), and estimated brain age of 4 normal subjects. And, we estimated brain age of 4 abnormal subjects to evaluate the detection accuracy of brain development abnormality. The results showed that the method found the differences of brain development for abnormal subjects.

Anterior cruciate ligament (ACL) injury causes knee joint instability, and effects on sports performance. Therefore, ACL reconstruction is essential to keep their high performance. It is well known that the outcome of ACL reconstruction is strongly related to the placement and orientation of the bone tunnel. Therefore, optimization of tunnel drilling technique is an important factor to obtain satisfactory surgical results. Current procedure relies on arthroscopic evaluation and there is a risk of damaging arteries and ligaments during surgery. The damages may reduce the accuracy and reproducibility of ACL reconstruction. As a postoperative evaluation method, a quadrant method has been used to evaluate the placement and orientation of the bone tunnel in X-ray radiography. This study proposes a computer-aided surgical planning system for evaluating ACL insertion site and orientation using magnetic resonance (MR) images. We first introduce MR image based the quadrant method to determine the ACL insertion site for preoperative patients. It also evaluates the 3-D spatial relationship between the planning femoral drilling hole and arteries around the femoral condyle. This system has been applied to ACL injured patients, it may increase the accuracy and reproducibility of ACL bone tunnel, and it can evaluate a risk of damaging the surrounding arteries and ligaments. (C) 2015 The Authors. Published by Elsevier B.V.

Immune cells deeply affect human health, however, it has not been investigated well due to difficulty of observing immune cells in vivo. The recent study enables us to acquire single Macrophage, which is the representative cell of immune cells, in vivo using 11.7 T magnetic resonance imaging (MRI). To quantify the kinematics of macrophages, it requires detection and tracking macrophages in MR image. The kinematic analysis will help researchers to investigate the mechanism of autoimmune diseases. This paper proposes an automated single macrophage tracking method in mouse brain 11.7 T time-lapse MR images. The method detects macrophages using background subtraction algorithm, and tracks macrophages using the Hungarian algorithm. The results showed that the proposed method detected and tracked macrophages in MR images successfully.

To prevent lifestyle diseases, this paper studies disease prediction using periodical health checkup data, daily monitoring to maintain healthy condition, and early life disease detection with medical imaging. To analyse periodical health checkup data, three approaches are introduced. The first approach is based on fuzzy set. It converts all attributes of health checkup data into fuzzy degrees by defining fuzzy membership functions. It enables us to manipulate all attributes in the same scale. The second approach analyses relationships between attributes of specific health examination data to cope with lifestyle diseases. It uses self-organizing maps, and clarifies the relationships among hemoglobin A1c (HbA1c), glutamic-oxaloacetic transaminase, glutamic-pyruvic transaminase, gamma-glutamyl transpeptidase, and triglyceride. The third approach predicts HbA1c fluctuations using decision tree. If we can predict the fluctuation, we can extract knowledge about what element will trigger developing diabetes. Through our examination, BMI will be the largest influencer about HbA1c fluctuations. Daily understanding of own condition is the first step of maintaining our health. A MEMS-based small and flexible monitoring device has been developed by the ERATO Maenaka human-sensing fusion project. We propose a condition estimation method using the monitoring device and FNN-based condition estimation. Experimental results show that it is a promising method for condition understanding. Cerebral vascular disease is one of major lifestyle diseases, and is caused by cerebral aneurysms. To predict the diseases, we should analyse cerebral arteries and aneurysms using magnetic resonance angiography images. This paper introduces an automated analysis method for early detection of aneurysms.

To inspect prescription drugs with press-through package (PTP), we propose an automated inspection system which based on computer vision. In the proposed system, we capture PTP drugs and apply hierarchical identification consist of several weak classifiers. In this paper, we report several results of inspection experiments which distinguish about a thousand kinds of PTPs. As a result, we have achieved sufficient recognition rate and processing time.

To avoid unattended death of 'single elderly people', this paper proposes a discreet surveillance system. The proposed surveillance system use wireless communication device to sense living activities, because employing surveillance camera will cause invasions of privacy. In this paper we report some results of preliminary experiments, which clarify relationship between wireless signal fluctuation and several living activities sleeping, deskwork, walk-through and so on. As a result, we can show that our system can sense living activities from signal fluctuations. © 2014 IEEE.

To inspect prescription drugs with press-through package (PTP), we propose an automated inspection system which based on computer vision. In the proposed system, we capture PTP drugs and apply hierarchical identification consist of several weak classifiers. In this paper, we report several results of inspection experiments which distinguish about a thousand kinds of PTPs. As a result, we have achieved sufficient recognition rate and processing time.

In this paper, we propose a method preventing howling from occurring by inserting a large loss into the talking circuit. Hands free telecommunication systems are generally formed by applying an acoustic echo canceler to reducing the acoustic coupling between a loudspeaker and a microphone. However, the system possibly causes the howling when the echo path change deteriorated the reduction effect of the acoustic coupling. The occurrence of the howling can be prevented by inserting a large loss into the talking circuit. Oppositely, this insertion loses the advantage of the two-way communication that is fulfilled in telecommunication system. We propose a method distinguishing between the echo path change and the double-talk and thereby restricting the insertion of the large loss only after the echo path change. Finally, we verify, using computer simulation, that the proposed method operates the hands free telecommunication system stably.

To inspect prescription drugs with one-dose package (ODP), we propose an automated inspection system which is based on computer vision. In the proposed system, we capture both sides of drugs at once and apply hierarchical identification. In this paper, we report several results of inspection experiments, which distinguish 60 kinds of naked drugs. As a result, we have achieved sufficient recognition rate and discuss its robustness. © 2013 Copyright TSI® Press.

In this paper, we propose a step size control method capable of quickly canceling acoustic echo even when double talk continues from the echo path change. This method controls the step size by substituting the norm of the difference vector between the coefficient vectors of a main adaptive filter (Main-ADF) and a sub-adaptive filter (Sub-ADF) for the estimation error provided by the former. Actually, the number of taps of Sub-ADF is limited to a quarter of that of Main-ADF, and the larger step size than that applied to Main-ADF is given to Sub-ADF; accordingly the norm of the difference vector quickly approximates to the estimation error. The estimation speed can be improved by utilizing the norm of the difference vector for the step size control in Main-ADF. We show using speech signals that in single talk the proposed method can provide almost the same estimation speed as the method whose step size is fixed at the optimum one and verify that even in double talk the estimation error, quickly decreases.

In this paper, we propose a method for repeatedly updating the coefficients of the feedback control filter used for canceling a feedback path from a loudspeaker to a noise detection microphone. The coefficients are usually estimated by feeding a sequence of extra noise to the loudspeaker. The proposed method can automatically estimate the coefficients without feeding the extra noise. We have already presented a method, called simultaneous equations method, based on a different principle from the filtered-x algorithm. The simultaneous equations method automatically estimates the coefficients of the noise control filter cancelling a target noise, called primary noise, without pre-identifying a secondary path from the loudspeaker to an error microphone. The simultaneous equations method can also derive the modified secondary path involving the feedback path and the primary path from the noise detection microphone to the error microphone. The proposed method estimates, utilizing the modified secondary path, the coefficients of the feedback control filter. In this paper, we show the updating process of the coefficients of the feedback control filter, and verify, using the impulse responses of practical acoustic paths, that the proposed method can estimate the feedback path and can cancel the primary noise without any pre-operations. Copyright© (2013) by Austrian Noise Abatement Association (OAL).

To inspect prescription drugs with press-through package (PTP), we propose an automated inspection system which based on computer vision. In the proposed system, we capture PTP drugs and apply hierarchical identification consist of several weak classifiers. In this paper, we report several results of inspection experiments which distinguish about a thousand kinds of PTPs. As a result, we have achieved sufficient recognition rate and processing time.

We're investigating an evaluation of safe driving by using computer vision technique to give drivers some indicators of their safe driving objectively. We previously reported that an estimation method of self-vehicle speed using feature points on road planes. And we also proposed an evaluation method of driving behavior using road-markings and stop time of self-vehicle. In this paper, we estimate stop time of self-vehicle with these 2 methods. By several experiments, we compare them and discuss which better is.

To recognize objects within narrow categories, it is important to extract effective features from small number of training samples. In this paper, first we discuss several depth features to improve object recognition accuracy. After that, we also discuss feature dimension reduction when we have insufficient training samples.

In this paper, we propose a normalization method dividing the gradient vector by the sum of the diagonal and two adjoining elements of the matrix expressing the correlation between the components of the discrete Fourier transform (DFT) of the reference signal used for the identification of unknown system. The proposed method can thereby improve the estimation speed of coefficients of adaptive filter.

In this paper, we propose a method capable of shortening the distance from a noise detection microphone to a loudspeaker in active noise control system with non-minimum phase secondary path. The distance can be basically shortened by forming the noise control filter, which produces the secondary noise provided by the loudspeaker, with the cascade connection of a non-recursive filter and a recursive filter. The output of the recursive filter, however, diverges even when the secondary path includes only a minimum phase component. In this paper, we prevent the divergence by utilizing MINT (multi-input/output inverse theorem) method increasing the number of secondary paths than that of primary paths. MINT method, however, requires a large scale inverse matrix operation, which increases the processing cost. We hence propose a method reducing the processing cost. Actually, MINT method has only to be applied to the non-minimum phase components of the secondary paths. We hence extract the non-minimum phase components and then apply MINT method only to those. The order of the inverse matrix thereby decreases and the processing cost can be reduced. We finally show a simulation result demonstrating that the proposed method successfully works.

In this paper, we propose a method shortening the distance from a noise detection microphone to a loudspeaker in active noise control system with non-minimum phase secondary path. The distance can be basically shortened by forming the noise control filter, which produces the secondary noise provided by the loudspeaker, with the cascade connection of non-recursive and recursive filters. The output of the recursive filter, however, diverges when the secondary path includes a minimum phase component. We hence prevent the divergence by extracting the non-minimum phase components and then apply MINT method only to them. We finally show a simulation result demonstrating that the proposed method successfully works.

We propose a two-channel adaptive algorithm efficiently utilizing independent components slightly involved in reference signals. In active noise control systems, the conventional methods of adding independent components to the reference signals or deliberately changing the correlation between the reference signals are inapplicable, because the reference signals are only detected by noise detection microphones. We have hence presented an adaptive algorithm that does not require the addition of the independent components or the deliberate change to the reference signals. However, this algorithm has three drawbacks. The first is that one of the convergence conditions is described by unmeasurable signals. The second is that pre-estimation of the reference signal power ratio is required. The third is that unknown acoustic paths cannot be identified when the reference signals have high cross-correlation between different sample times. In this work, we derive a new algorithm without the drawbacks and verify, using computer simulations, that the derived algorithm can successfully identify the unknown acoustic paths. The two-channel active noise control system using the simultaneous equations method is thereby expected to provide a high noise reduction effect. © 2012 The Acoustical Society of Japan.

This paper estimates some vehicle speeds from video sequences which was shot by in-vehicle camera. By giving some structural information of target object, we can estimate oncoming vehicle speed accurately. Some experimental results show that our method can estimate other moving vehicles with enough accuracy for investigating traffic accidents.