Shoko Miyauchi

Shoko Miyauchi

Kyushu University

H-index: 6

Asia-Japan

About Shoko Miyauchi

Shoko Miyauchi, With an exceptional h-index of 6 and a recent h-index of 5 (since 2020), a distinguished researcher at Kyushu University, specializes in the field of Statistical Shape Model, Mapping, Correspondence, Medical image processing.

His recent articles reflect a diverse array of research interests and contributions to the field:

The Kumagai Method Utilizing the Pigeon Bottle Feeder with a Long Nipple: A Descriptive Study for the Development of Feeding Techniques for Children with Cleft Lip and/or Palate

FusionNet: A Frame Interpolation Network for 4D Heart Models

Personalized Federated Learning for Institutional Prediction Model using Electronic Health Records: A Covariate Adjustment Approach

Multiple Instance Learning による大腸病理画像からの癌再発予測システムの構築

Isomorphic mesh generation from point clouds with multilayer perceptrons

Development of AR training systems for Humanitude dementia care

Esophageal Tumor Segmentation in Endoscopic Images by Deep Learning

Automatic electron hologram acquisition of catalyst nanoparticles using particle detection with image processing and machine learning

Shoko Miyauchi Information

University

Kyushu University

Position

___

Citations(all)

89

Citations(since 2020)

65

Cited By

29

hIndex(all)

6

hIndex(since 2020)

5

i10Index(all)

5

i10Index(since 2020)

3

Email

University Profile Page

Kyushu University

Shoko Miyauchi Skills & Research Interests

Statistical Shape Model

Mapping

Correspondence

Medical image processing

Top articles of Shoko Miyauchi

The Kumagai Method Utilizing the Pigeon Bottle Feeder with a Long Nipple: A Descriptive Study for the Development of Feeding Techniques for Children with Cleft Lip and/or Palate

Authors

Shingo Ueki,Yukari Kumagai,Yumi Hirai,Eri Nagatomo,Shoko Miyauchi,Takuro Inoue,Qi An,Junko Miyata

Journal

Children

Published Date

2024/3/19

We aimed to identify the steps involved in the Kumagai method—an experimental nursing procedure to feed children with cleft lip and/or palate, using a feeder with a long nipple. We conducted a descriptive study, enrolling five specialist nurses who have mastered the Kumagai method. Their approaches were examined using structured interviews. Moreover, the participants were asked to perform the sequence of actions involved in this method while describing each step. Therefore, we were able to explore the Kumagai method in depth and step-by-step, including the following aspects: correct infant posture; correct feeding bottle holding position; nipple insertion into the child’s mouth; and feeding process initiation, maintenance, and termination. Each step comprises several clinically relevant aspects aimed at encouraging the infant to suck with a closed mouth and stimulating chokubo-zui, i.e., simulation of the natural tongue movement during breastfeeding in children without a cleft palate. In conclusion, when performed correctly, the Kumagai method improves feeding efficiency in children with cleft lip and/or palate. Feeders with long nipples are rarely used in clinical practice; the Kumagai method might popularize their use, thereby improving the management of feeding practices for children with cleft lip and/or palate.

FusionNet: A Frame Interpolation Network for 4D Heart Models

Authors

Chujie Chang,Shoko Miyauchi,Ken’ichi Morooka,Ryo Kurazume,Oscar Martinez Mozos

Published Date

2023/10/8

Cardiac magnetic resonance (CMR) imaging is widely used to visualise cardiac motion and diagnose heart disease. However, standard CMR imaging requires patients to lie still in a confined space inside a loud machine for 40–60 min, which increases patient discomfort. In addition, shorter scan times decrease either or both the temporal and spatial resolutions of cardiac motion, and thus, the diagnostic accuracy of the procedure. Of these, we focus on reduced temporal resolution and propose a neural network called FusionNet to obtain four-dimensional (4D) cardiac motion with high temporal resolution from CMR images captured in a short period of time. The model estimates intermediate 3D heart shapes based on adjacent shapes. The results of an experimental evaluation of the proposed FusionNet model showed that it achieved a performance of over 0.897 in terms of the Dice coefficient, confirming that it can …

Personalized Federated Learning for Institutional Prediction Model using Electronic Health Records: A Covariate Adjustment Approach

Authors

Shinji Tarumi,Mayumi Suzuki,Hanae Yoshida,Shoko Miyauchi,Ryo Kurazume

Published Date

2023/7/24

Federated learning (FL) has attracted attention as a technology that allows multiple medical institutions to collaborate on AI without disclosing each other's patient data. However, FL has the challenge of being unable to robustly learn when the data of participating clients is non-independently and non-identically distributed (Non-IID). Personalized Federated Learning (PFL), which constructs a personalized model for each client, has been proposed as a solution to this problem. However, conventional PFL methods do not ensure the interpretability of personalization, specifically, the identification of which data samples are contributed to each personalized learning, which is important for AI in medical applications. In this study, we propose a novel PFL framework, Federated Adjustment of Covariate (FedCov), which acquires a propensity score model representing the covariate shift among clients through prior FL, then …

Multiple Instance Learning による大腸病理画像からの癌再発予測システムの構築

Authors

大森一輝, 諸岡健一, 中西良太, 宮内翔子, 沖英次, 吉住朋晴

Journal

電子情報通信学会技術研究報告; 信学技報

Published Date

2022/5/12

抄録 (和) 大腸癌は, 再発すると死亡のリスクが非常に高くなるため, 再発リスクが高いと考えられる場合は術後に化学療法を行う. しかし, 現在までに決定的な再発リスクの判断根拠は見つかっていないため, 医師が癌のステージや腫瘍の深さなどから総合的に判断し, 治療方針を決定している. そこで, 本研究では, 大腸標本の Whole Slide Image, およびそれに含まれる大腸癌領域を入力とし, Multiple Instance Learing を用いて, 再発リスクの予測とその診断根拠を提示するシステムを提案する.

Isomorphic mesh generation from point clouds with multilayer perceptrons

Authors

Shoko Miyauchi,Ken'ichi Morooka,Ryo Kurazume

Journal

arXiv preprint arXiv:2210.14157

Published Date

2022/10/21

We propose a new neural network, called isomorphic mesh generator (iMG), which generates isomorphic meshes from point clouds containing noise and missing parts. Isomorphic meshes of arbitrary objects have a unified mesh structure even though the objects belong to different classes. This unified representation enables surface models to be handled by DNNs. Moreover, the unified mesh structure of isomorphic meshes enables the same process to be applied to all isomorphic meshes; although in the case of general mesh models, we need to consider the processes depending on their mesh structures. Therefore, the use of isomorphic meshes leads to efficient memory usage and calculation time compared with general mesh models. As iMG is a data-free method, preparing any point clouds as training data in advance is unnecessary, except a point cloud of the target object used as the input data of iMG. Additionally, iMG outputs an isomorphic mesh obtained by mapping a reference mesh to a given input point cloud. To estimate the mapping function stably, we introduce a step-by-step mapping strategy. This strategy achieves a flexible deformation while maintaining the structure of the reference mesh. From simulation and experiments using a mobile phone, we confirmed that iMG can generate isomorphic meshes of given objects reliably even when the input point cloud includes noise and missing parts.

Development of AR training systems for Humanitude dementia care

Authors

Ryo Kurazume,Tomoki Hiramatsu,Masaya Kamei,Daiji Inoue,Akihiro Kawamura,Shoko Miyauchi,Qi An

Journal

Advanced Robotics

Published Date

2022/4/3

This paper presents a comparative study of two AR training systems for Humanitude dementia care. Humanitude is a multimodal comprehensive care methodology for patients with dementia and has attracted considerable attention as a gentle and effective care technique. To provide an effective training system for Humanitude, we developed HEARTS 1, which realizes simultaneous sensing and interaction by combining a real training entity and augmented reality technology. However, training experiments using HEARTS 1 indicated that HEARTS 1 had certain weaknesses that should be addressed, namely, the usability of the AR display or the impression of the CG model. This paper presents a new prototype called HEARTS 2 consisting of Microsoft HoloLens 2 as well as realistic and animated CG models of older women. Psychological experiments were conducted to identify the difference in the usability of the AR …

Esophageal Tumor Segmentation in Endoscopic Images by Deep Learning

Authors

Zehao Li,Ken'ichi Morooka,Yuho Ebata,Hirofumi Hasuda,Shoko Miyauchi,Ota Mitsuhiko

Journal

IEICE Technical Report; IEICE Tech. Rep.

Published Date

2022/9/8

(in English) Esophageal cancer is often asymptomatic at early stage. It progresses rapidly and can invade surrounding tissues. The esophagus is surrounded by vital organs. Treatment becomes difficult if these organs invaded by esophageal cancer. Therefore, early diagnosis of esophageal cancer is necessary. In this study, we constructed a system for extracting tumor regions from NBI endoscopic images obtained by Narrow Band Imaging (NBI), and detected tumor regions with an accuracy of 83.5%.

Automatic electron hologram acquisition of catalyst nanoparticles using particle detection with image processing and machine learning

Authors

Fumiaki Ichihashi,Akira Koyama,Tetsuya Akashi,Shoko Miyauchi,Ken'ichi Morooka,Hajime Hojo,Hisahiro Einaga,Yoshio Takahashi,Toshiaki Tanigaki,Hiroyuki Shinada,Yasukazu Murakami

Journal

Applied Physics Letters

Published Date

2022/2/7

To enable better statistical analysis of catalyst nanoparticles by high-resolution electron holography, we improved the particle detection accuracy of our previously developed automated hologram acquisition system by using an image classifier trained with machine learning. The detection accuracy of 83% was achieved with the small training data of just 232 images showing nanoparticles by utilizing transfer learning based on VGG16 to train the image classifier. Although the construction of training data generally requires much effort, the time needed to select the training data candidates was significantly shortened by utilizing a pattern matching technique. Experimental results showed that the high-resolution hologram acquisition efficiency was improved by factors of about 100 and 6 compared to a scan method and a patternmatching-only method, respectively.

深層学習による NBI 内視鏡画像を用いた上部消化管腫瘍抽出

Authors

李澤昊, 諸岡健一, 江端由穂, 蓮田博文, 宮内翔子, 太田光彦

Journal

電子情報通信学会技術研究報告; 信学技報

Published Date

2022/9/8

抄録 (和) 食道癌は, 発症早期が無症状で進行が早く, また周囲に浸潤しやすい癌である. 食道の周囲には重要臓器に取り囲まれており, これらの臓器に食道癌が浸潤すると治療が困難になる. したがって, 食道癌の早期診断は重要である. 本研究では, Narrow Band Imaging (NBI) によって得られた NBI 内視鏡画像から腫瘍領域を抽出するシステムを構築し, 83.5% の精度で腫瘍領域を検出できた.(英) Esophageal cancer is often asymptomatic at early stage. It progresses rapidly and can invade surrounding tissues. The esophagus is surrounded by vital organs. Treatment becomes difficult if these organs invaded by esophageal cancer. Therefore, early diagnosis of esophageal cancer is necessary. In this study, we constructed a system for extracting tumor regions from NBI endoscopic images obtained by Narrow Band Imaging (NBI), and detected tumor regions with an accuracy of 83.5%.

A Computer-Aided Support System for Deep Brain Stimulation by Multidisciplinary Brain Atlas Database

Authors

Ken’ichi Morooka,Shoko Miyauchi,Yasushi Miyagi

Journal

Multidisciplinary Computational Anatomy: Toward Integration of Artificial Intelligence with MCA-based Medicine

Published Date

2022

Our research group has been constructing a 3D digital atlas of a Japanese brain (Fukuda et al, Neurosci Res. 67:260–265, 2010, Miyauchi et al. Proc. Computer Assisted Radiology and Surgery. 232, 2017). The purpose of our project is to develop a support system for Deep Brain Stimulation using the brain atlas. Practically, the support system estimates a patient brain atlas by using a multidisciplinary brain atlas database based on our 3D brain atlas. To achieve this, we have been developing two fundamental techniques. The first is to deform our brain atlas nonlinearly to estimate a reliable atlas of a patient with acceptable accuracy for practical medical cases. The second is to determine the correspondences among the atlases of different individuals to construct the brain atlas database.

患者の心臓動的形状とメタデータを用いた虚血性心疾患診断システムの構築

Authors

宮内翔子, 諸岡健一, 倉爪亮

Journal

電子情報通信学会技術研究報告; 信学技報

Published Date

2022/7/1

抄録 (和) 心臓 MR 画像から抽出した心臓動的形状と, 患者のメタデータを用いて, 虚血性心疾患と健常者を分類する虚血性心疾患診断システムを提案する. 提案システムのメタデータとしては, BMI や血圧, 喫煙の有無など, 虚血性心疾患の危険因子と関連の高いものを用いる. 健常者 100 名分と虚血性心疾患患者 100 名分を用いた実験を行い, 心臓動的形状のみやメタデータのみを提案システムに入力した場合よりも, 心臓動的形状とメタデータ両方を入力し

陰関数表現を用いた同一構造を持つ 3 次元物体メッシュモデル生成法の構築

Authors

板谷響, 宮内翔子, 諸岡健一

Journal

研究報告コンピュータグラフィックスとビジュアル情報学 (CG)

Published Date

2022/11/11

本研究では, 符号付き距離場で表されたクラスの異なる 3 次元物体を, 同一のメッシュ構造で記述する方法を提案する. まず, クラスごとに, それに属する物体間で共通する形状特徴を有するテンプレート形状を, 陰関数表現を用いて学習する. 次に, 学習した全てのクラスのテンプレート形状を, 事前に用意した基準メッシュモデルのメッシュ構造で記述する. 最後に, 各クラスのテンプレート形状から, そのクラスの 3 次元物体への逆写像を行うことで, 全ての 3 次元物体を共通のメッシュ構造で記述する. 4 クラスの 3 次元物体を用いてメッシュモデルの生成を行った結果, 従来手法と比較して, 計算コストを低減しつつ, 高精度な形状復元が可能であることを確認した. また, 提案手法で生成したメッシュモデルを用いて深層学習による 3 次元物体識別を行った結果, 従来手法と同等以上の精度で 3 次元物体を識別できることを確認した.

Artificial intelligence for segmentation of bladder tumor cystoscopic images performed by U-Net with dilated convolution

Authors

Jun Mutaguchi,Ken'ichi Morooka,Satoshi Kobayashi,Aiko Umehara,Shoko Miyauchi,Fumio Kinoshita,Junichi Inokuchi,Yoshinao Oda,Ryo Kurazume,Masatoshi Eto

Journal

Journal of Endourology

Published Date

2022/6/1

Background: Early intravesical recurrence after transurethral resection of bladder tumors (TURBT) is often caused by overlooking of tumors during TURBT. Although narrow-band imaging and photodynamic diagnosis were developed to detect more tumors than conventional white-light imaging, the accuracy of these systems has been subjective, along with poor reproducibility due to their dependence on the physician's experience and skills. To create an objective and reproducible diagnosing system, we aimed at assessing the utility of artificial intelligence (AI) with Dilated U-Net to reduce the risk of overlooked bladder tumors when compared with the conventional AI system, termed U-Net. Materials and Methods: We retrospectively obtained cystoscopic images by converting videos obtained from 120 patients who underwent TURBT into 1790 cystoscopic images. The Dilated U-Net, which is an extension of the …

3 次元心臓モデルのためのフレーム補間手法の構築

Authors

宮内翔子, 諸岡健一, 倉爪亮

Journal

研究報告コンピュータビジョンとイメージメディア (CVIM)

Published Date

2022/11/11

論文抄録医用画像の一つである心臓 MRI (Magnetic resonance imaging) は, 心疾患の診断のために広く用いられている. しかし, 心臓 MRI の撮像時間は約 50 分と長く, 撮像コストや患者への負担が大きい. 撮像時間を短くした場合, 得られる画像のフレームレートが低下し, そこから観察される心臓の動きの時間分解能も低下する. 心疾患の診断では, 心臓の動きの観察が重要となるため, 時間分解能の低下により, 診断精度が低下する恐れがある. そこで, 我々は, 短時間で撮像された時間分解能の低い心臓 MRI から, 高い時間分解能の心臓の動きを推定するための生成ネットワークを提案する. 提案ネットワークを用いることで, 従来の生成モデルを用いた場合よりも高い精度で, 心臓の動きの時間分解能を向上できることを確認した.

Automatic Hologram Acquisition of Pt Catalyst Nanoparticles on TiO2 Using Particle Detection with Image Processing and AI Classification

Authors

Fumiaki Ichihashi,Akira Koyama,Tetsuya Akashi,Shoko Miyauchi,Ken'ichi Morooka,Hajime Hojo,Hisahiro Einaga,Toshiaki Tanigaki,Hiroyuki Shinada,Yasukazu Murakami

Journal

Microscopy and Microanalysis

Published Date

2021/8

Catalysts of metal nanoparticles have attracted much attention for various applications, such as air purification, hydrogen production, and fuel cells, and clarifying the mechanism of the catalytic activity is important. Electron holography is a microscopic method that detects the phase change that an electron wave undergoes when it propagates in an electromagnetic potential. The mechanism of catalytic activities is revealed by measuring the electric field around nanoparticles using electron holography. However, a measurement with a high signal-to-noise (SN) ratio is required because the amount of phase change due to the electric field around the nanoparticles is very small. The phase measurement sensitivity should be improved by integrating and averaging a large number of phase images reconstructed from holograms. Therefore, we have developed an automatic acquisition system that detects particle positions …

MICCAI 2020 参加報告

Authors

斉藤篤, 小田昌宏, 大竹義人, 花岡昇平, 諸岡健一, 宮内翔子, 増谷佳孝, 申忱, 森健策

Journal

電子情報通信学会技術研究報告; 信学技報

Published Date

2021/3/8

抄録 (和) 本稿では, MICCAI 2020 の本会議セッション及びワークショップの概要を紹介し, 特に興味深い報告について内容を解説する.

A deep learning-based method for predicting volumes of nasopharyngeal carcinoma for adaptive radiation therapy treatment

Authors

Bilel Daoud,Ken'ichi Morooka,Shoko Miyauchi,Ryo Kurazume,Wafa Mnejja,Leila Farhat,Jamel Daoud

Published Date

2021/1/10

This paper presents a new system for predicting the spatial change of Nasopharyngeal carcinoma(NPC) and organ-at-risks (OARs) volumes over the course of the radiation therapy (RT) treatment for facilitating the workflow of adaptive radiotherapy. The proposed system, called “Tumor Evolution Prediction (TEP-Net)”, predicts the spatial distributions of NPC and 5 OARs, separately, in response to RT in the coming week, week n. Here, TEP-Net has (n-1)-inputs that are week 1 to week n-1 of CT axial, coronal or sagittal images acquired once the patient complete the planned RT treatment of the corresponding week. As a result, three predicted results of each target region are obtained from the three-view CT images. To determine the final prediction of NPC and 5 OARs, two integration methods, weighted fully connected layers and weighted voting methods, are introduced. From the experiments using weekly CT …

Motion generation by learning relationship between object shapes and human motions

Authors

Tokuo Tsuji,Sho Tajima,Yosuke Suzuki,Tetsuyou Watanabe,Shoko Miyauchi,Ken’Ichi Morooka,Kensuke Harada,Hiroaki Seki

Published Date

2021

This paper presents a method for planning a robot motion of daily tasks by learning the relationship between object shapes and human motions. Robots are required to be able to deal with multifarious objects in various categories. However, it is difficult for robots to plan motions automatically for performing a task because objects even in the same category have different shapes. In our method, the motions are estimated by learning the relationship between object shapes and human motions using linear regression analysis.

日本医用画像工学会 (JAMIT) 会長 尾川浩一日本医用画像工学会第 40 回大会 大会長 陣崎雅弘

Authors

宮内翔子

Journal

Medical Imaging Technology

Published Date

2021/1

編集後記 Toggle navigation J-STAGE home 資料・記事を探す 資料を探す:資料タイトルから 資料を探す:分野から 資料を探す:発行機関から 記事を探す データを探す(J-STAGE Data) J-STAGEについて J-STAGEの概要 各種サービス・機能 公開データ 利用規約・ポリシー 新規登載の申し込み ニュース &PR お知らせ一覧 リリースノート メンテナンス情報 イベント情報 J-STAGEニュース 特集コンテンツ 各種広報媒体 サポート J-STAGE登載機関用コンテンツ 登載ガイドライン・マニュアル 閲覧者向け ヘルプ 動作確認済みブラウザ FAQ お問い合わせ サイトマップ サインイン カート JA English 日本 語 資料・記事を探す 資料を探す:資料タイトルから 資料を探す:分野から 資料を探す:発行機関から 記事を探す データを探す(J-STAGE Data) J-STAGEについて J-STAGEの概要 各種サービス・機能 公開データ 利用規約・ポリシー 新規登載の申し込み ニュース&PR お知らせ一覧 リリースノート メンテナンス情報 イベント情報 J-STAGEニュース 特集コンテンツ 各種広報媒体 サポート J-STAGE…

Analysis of TEM images of metallic nanoparticles using convolutional neural networks and transfer learning

Authors

Akira Koyama,Shoko Miyauchi,Ken'ichi Morooka,Hajime Hojo,Hisahiro Einaga,Yasukazu Murakami

Journal

Journal of Magnetism and Magnetic Materials

Published Date

2021/11/15

Convolutional neural networks (CNNs) pretrained by transfer learning were applied to the analysis of transmission electron microscopy (TEM) images of nanoparticles. Specifically, TEM images of non-magnetic Pt nanoparticles dispersed on a thin TiO 2 crystal foil were classified using CNNs. Although the number of learning data (50≤ N≤ 350) was several orders of magnitude smaller than the quantities normally employed in conventional CNN analyses, the present CNN model was able to carry out image classification with 94% accuracy (average of 25 results) after the convolutional layers were pretrained by transfer learning and fine tuning. This method represents a promising tool for TEM studies of both non-magnetic and magnetic nanoparticles which make emergence of rich material functions.

See List of Professors in Shoko Miyauchi University(Kyushu University)

Shoko Miyauchi FAQs

What is Shoko Miyauchi's h-index at Kyushu University?

The h-index of Shoko Miyauchi has been 5 since 2020 and 6 in total.

What are Shoko Miyauchi's top articles?

The articles with the titles of

The Kumagai Method Utilizing the Pigeon Bottle Feeder with a Long Nipple: A Descriptive Study for the Development of Feeding Techniques for Children with Cleft Lip and/or Palate

FusionNet: A Frame Interpolation Network for 4D Heart Models

Personalized Federated Learning for Institutional Prediction Model using Electronic Health Records: A Covariate Adjustment Approach

Multiple Instance Learning による大腸病理画像からの癌再発予測システムの構築

Isomorphic mesh generation from point clouds with multilayer perceptrons

Development of AR training systems for Humanitude dementia care

Esophageal Tumor Segmentation in Endoscopic Images by Deep Learning

Automatic electron hologram acquisition of catalyst nanoparticles using particle detection with image processing and machine learning

...

are the top articles of Shoko Miyauchi at Kyushu University.

What are Shoko Miyauchi's research interests?

The research interests of Shoko Miyauchi are: Statistical Shape Model, Mapping, Correspondence, Medical image processing

What is Shoko Miyauchi's total number of citations?

Shoko Miyauchi has 89 citations in total.

    academic-engine

    Useful Links