博士（工学）, 東京工業大学

PhD, Tokyo Institute of Technology

Natural sciences, Applied mathematics and statistics

Natural sciences, Basic mathematics

Informatics, Intelligent informatics

Informatics, Statistical science

Informatics, Mathematical informatics

Oct. 2014 - Present
The University of Electro- Communications, Graduate School of Information Systems, Assistant Professor

Apr. 2014 - Present
Tohoku University, Division of Biomedical Information Analysis, Department of Integrative Genomics, Tohoku Medical Megabank Organization, Part-time Lecturer

Apr. 2013 - Sep. 2014
The Institute of Statistical Mathematics, Research Center for Statistical Machine Learning, Project Assistant Professor

Jan. 2012 - Mar. 2013
The Institute of Statistical Mathematics, Research Center for Statistical Machine Learning, Project Researcher

Apr. 2011 - Dec. 2011
The Institute of Statistical Mathematics, Research Innovation Center, Project Researcher

Apr. 2010 - Mar. 2011
RIKEN Brain Science Institute (RIKEN BSI), Integrated Theoretical Neuroscience, Postdoctoral Fellow

Oct. 2009 - Mar. 2010
Tokyo Institute of Technology, Precision and Intelligence Laboratory (P&I Lab), Research fellow (PD) of Japan Society for the Promotion of Science (JSPS)

Apr. 2007 - Sep. 2009
Japan Society for the Promotion of Science (JSPS), Research fellow (DC1)

Oct. 2006 - Sep. 2009
Tokyo Institute of Technology, Department of Computational Intelligence and Systems Science, Doctor of Engineering

Apr. 2005 - Sep. 2006
Tokyo Institute of Technology, Department of Computational Intelligence and Systems Science, Master

Apr. 2001 - Mar. 2005
Keio University, Department of Physics, Bachelor

Apr. 1997 - Mar. 2000
私立桐蔭学園高等学校

01 Oct. 2006 - 30 Sep. 2009
東京工業大学, 大学院総合理工学研究科, 知能システム科学専攻

01 Apr. 2005 - 30 Sep. 2006
東京工業大学, 大学院総合理工学研究科, 知能システム科学専攻

01 Apr. 2001 - 31 Mar. 2005
慶應義塾大学, 理工学部, 物理学科

01 Apr. 1997 - 31 Mar. 2000
桐蔭学園高等学校

Feb. 2019
動体追跡システムに関する研究
株式会社TOA開発室研究奨励賞, 大塚研秀;近藤亮祐;冨田恭平;西山悠;小木曽公尚;饗庭絵里子;小泉憲裕
Others

Jan. 2019
古野電気株式会社技術研究所所長賞, 近藤亮祐;今泉飛翔;西山悠;小泉憲裕
Others

Mar. 2012
部分観測マルコフ決定過程ベルマン方程式のカーネル化
IBISML 2011年度 研究会賞ファイナリスト, 西山悠;Abdeslam Boularias;Arthur Gretton;福水健次

Mar. 2009
IEEE Computational Intelligence Society Japan (CISJ), 受賞形態: 個人、国内外区分: 国内の賞、専門分野: 計算論的神経科学
Young Researcher Award, 2008.
ベーテ自由エネルギーに対するCCCPアルゴリズムの拡張, 西山悠

Image Search Strategy via Visual Servoing for Robotic Kidney Ultrasound Imaging
Takumi Fujibayashi; Norihiro Koizumi; Yu Nishiyama; Jiayi Zhou; Hiroyuki Tsukihara; Kiyoshi Yoshinaka; Ryosuke Tsumura
Journal of Robotics and Mechatronics, 20 Oct. 2023, Peer-reviwed
Scientific journal

Rib region detection for scanning path planning for fully automated robotic abdominal ultrasonography.
Koudai Okuzaki; Norihiro Koizumi; Kiyoshi Yoshinaka; Yu Nishiyama; Jiayi Zhou; Ryosuke Tsumura
International journal of computer assisted radiology and surgery, 03 Oct. 2023, Peer-reviwed, True, PURPOSE: Scanning path planning is an essential technology for fully automated ultrasound (US) robotics. During biliary scanning, the subcostal boundary is critical body surface landmarks for scanning path planning but are often invisible, depending on the individual. This study developed a method of estimating the rib region for scanning path planning toward fully automated robotic US systems. METHODS: We proposed a method for determining the rib region using RGB-D images and respiratory variation. We hypothesized that detecting the rib region would be possible based on changes in body surface position due to breathing. We generated a depth difference image by finding the difference between the depth image taken at the resting inspiratory position and the depth image taken at the maximum inspiratory position, which clearly shows the rib region. The boundary position of the subcostal was then determined by applying training using the YOLOv5 object detection model to this depth difference image. RESULTS: In the experiments with healthy subjects, the proposed method of rib detection using the depth difference image marked an intersection over union (IoU) of 0.951 and average confidence of 0.77. The average error between the ground truth and predicted positions was 16.5 mm in 3D space. The results were superior to rib detection using only the RGB image. CONCLUSION: The proposed depth difference imaging method, which measures respiratory variation, was able to accurately estimate the rib region without contact and physician intervention. It will be useful for planning the scan path during the biliary imaging.
Scientific journal, English

Study on method of organ section retention and tracking through deep learning in automated diagnostic and therapeutic robotics.
Takumi Fujibayashi; Norihiro Koizumi; Yu Nishiyama; Yusuke Watanabe; Jiayi Zhou; Momoko Matsuyama; Miyu Yamada; Ryosuke Tsumura; Kiyoshi Yoshinaka; Naoki Matsumoto; Hiroyuki Tsukihara; Kazushi Numata
International journal of computer assisted radiology and surgery, 30 May 2023, Peer-reviwed, True, PURPOSE: In high-intensity focused ultrasound (HIFU) treatment of the kidney and liver, tracking the organs is essential because respiratory motions make continuous cauterization of the affected area difficult and may cause damage to other parts of the body. In this study, we propose a tracking system for rotational scanning, and propose and evaluate a method for estimating the angles of organs in ultrasound images. METHODS: We proposed AEMA, AEMAD, and AEMAD++ as methods for estimating the angles of organs in ultrasound images, using RUDS and a phantom to acquire 90-degree images of a kidney from the long-axis image to the short-axis image as a data set. Six datasets were used, with five for preliminary preparation and one for testing, while the initial position was shifted by 2 mm in the contralateral axis direction. The test data set was evaluated by estimating the angle using each method. RESULTS: The accuracy and processing speed of angle estimation for AEMA, AEMAD, and AEMAD++ were 23.8% and 0.33 FPS for AEMAD, 32.0% and 0.56 FPS for AEMAD, and 29.5% and 3.20 FPS for AEMAD++, with tolerance of ± 2.5 degrees. AEMAD++ offered the best speed and accuracy. CONCLUSION: In the phantom experiment, AEMAD++ showed the effectiveness of tracking the long-axis image of the kidney in rotational scanning. In the future, we will add either the area of surrounding organs or the internal structure of the kidney as a new feature to validate the results.
Scientific journal, English

超音波画像処理と位置情報を融合した新技術展開 深層学習に基づくロボティック超音波診断支援システムの開発
Zhou Jiayi; 小泉 憲裕; 西山 悠; 津村 遼介; 葭仲 潔; 松本 直樹; 小川 眞広; 沼田 和司
超音波医学, (公社)日本超音波医学会, 50, Suppl., S507-S507, Apr. 2023
Japanese

A VS ultrasound diagnostic system with kidney image evaluation functions
Jiayi Zhou; Norihiro Koizumi; Yu Nishiyama; Kiminao Kogiso; Tomohiro Ishikawa; Kento Kobayashi; Yusuke Watanabe; Takumi Fujibayashi; Miyu Yamada; Momoko Matsuyama; Hiroyuki Tsukihara; Ryosuke Tsumura; Kiyoshi Yoshinaka; Naoki Matsumoto; Masahiro Ogawa; Hideyo Miyazaki; Kazushi Numata; Hidetoshi Nagaoka; Toshiyuki Iwai; Hideyuki Iijima
International Journal of Computer Assisted Radiology and Surgery, 18, 2, 227-246, Feb. 2023, Peer-reviwed, True
Scientific journal, English

An automatic judgment method of swelling or atrophy of organs for ultrasound diagnosis
Miyu Yamada; Ryosuke Tsumura; Norihiro Koizumi; Kiyoshi Yoshinaka; Yu Nishiyama; Naoki Matsumoto
GCCE 2022 - 2022 IEEE 11th Global Conference on Consumer Electronics, 757-758, Oct. 2022
International conference proceedings

An avoiding overlap method between acoustic shadow and organ for automated ultrasound diagnosis and treatment
Momoko Matsuyama; Norihiro Koizumi; Yu Nishiyama; Ryosuke Tsumura; Hiroyuki Tsukihara; Kazushi Numata
GCCE 2022 - 2022 IEEE 11th Global Conference on Consumer Electronics, 746-747, Oct. 2022
International conference proceedings

A study on the same cross-sectional tracking method using AEMADP++ based on YOLACT++ for automated diagnostic and therapeutic robots∗
Takumi Fujibayashi; Norihiro Koizumi; Yu Nishiyama; Ryosuke Tsumura; Kiyoshi Yoshinaka; Hiroyuki Tsukihara
GCCE 2022 - 2022 IEEE 11th Global Conference on Consumer Electronics, 744-745, Oct. 2022
International conference proceedings

Development of a VS ultrasound diagnostic system with image evaluation functions
Jiayi Zhou; Norihiro Koizumi; Yu Nishiyama; Ryosuke Tsumura; Hiroyuki Tsukihara; Naoki Matsumoto
GCCE 2022 - 2022 IEEE 11th Global Conference on Consumer Electronics, 699-700, Oct. 2022
International conference proceedings

Artificial intelligence for distinguishment of hammering sound in total hip arthroplasty
Yasuhiro Homma; Shun Ito; Xu Zhuang; Tomonori Baba; Kazutoshi Fujibayashi; Kazuo Kaneko; Yu Nishiyama; Muneaki Ishijima
SCIENTIFIC REPORTS, 12, 1, 9826-9826, 14 Jun. 2022, Peer-reviwed, True
Scientific journal, English

診断画像適正度の評価のための深層学習を用いた臓器検出と診断画像欠損部の同定
桂木 嵐; 小泉 憲裕; 西山 悠; 山田 望結; 藤林 巧; 沼田 和司; 月原 弘之; 増崎 亮太; 松本 直樹; 小川 眞広
超音波医学, (公社)日本超音波医学会, 49, Suppl., S581-S581, Apr. 2022
Japanese

A novel complementation method of an acoustic shadow region utilizing a convolutional neural network for ultrasound-guided therapy
Momoko Matsuyama; Norihiro Koizumi; Akihide Otsuka; Kento Kobayashi; Shiho Yagasaki; Yusuke Watanabe; Jiayi Zhou; Yu Nishiyama; Naoki Matsumoto; Hiroyuki Tsukihara; Kazushi Numata
International Journal of Computer Assisted Radiology and Surgery, 17, 1, 107-119, Jan. 2022, Peer-reviwed, True
Scientific journal, English

動的輪郭モデルを用いた深層学習による2次元量音波診断画像からの3次元移動量推定
矢ヶ崎 詞穂; 小泉 憲裕; 西山 悠; 齋藤 僚介; 小川 眞広; 松本 直樹; 沼田 和司
日本コンピュータ外科学会誌, (一社)日本コンピュータ外科学会, 23, 4, 221-221, Nov. 2021
Japanese

Evaluation of ultrasonic fibrosis diagnostic system using convolutional network for ordinal regression
Ryosuke Saito; Norihiro Koizumi; Yu Nishiyama; Tsubasa Imaizumi; Kenta Kusahara; Shiho Yagasaki; Naoki Matsumoto; Ryota Masuzaki; Toshimi Takahashi; Masahiro Ogawa
International Journal of Computer Assisted Radiology and Surgery, 16, 11, 1969-1975, Nov. 2021, Peer-reviwed, True
Scientific journal, English

人工知能を用いた肝線維化の超音波診断システムの開発
齋藤 僚介; 小泉 憲裕; 西山 悠; 今泉 飛翔; 草原 健太; 矢ヶ崎 詩穂; 小川 眞広; 松本 直樹
超音波医学, (公社)日本超音波医学会, 48, Suppl., S652-S652, Apr. 2021
Japanese

深層学習を用いた2次元超音波画像からの3次元移動量推定
矢ヶ崎 詞穂; 小泉 憲裕; 西山 悠; 近藤 亮祐; 草原 健太; 齋藤 僚介; 小川 眞広; 松本 直樹; 沼田 和司
超音波医学, (公社)日本超音波医学会, 48, Suppl., S653-S653, Apr. 2021
Japanese

Estimating 3-dimensional liver motion using deep learning and 2-dimensional ultrasound images
Shiho Yagasaki; Norihiro Koizumi; Yu Nishiyama; Ryosuke Kondo; Tsubasa Imaizumi; Naoki Matsumoto; Masahiro Ogawa; Kazushi Numata
International Journal of Computer Assisted Radiology and Surgery, 15, 12, 1989-1995, Dec. 2020, Peer-reviwed, True
Scientific journal, English

Sagittal alignment in an MR-TRUS fusion biopsy using only the prostate contour in the axial image
Riki Igarasihi; Norihiro Koizumi; Yu Nishiyama; Kyohei Tomita; Yuka Shigenari; Sunao Shoji
ROBOMECH Journal, 7, 1, 01 Dec. 2020, Peer-reviwed
Scientific journal, English

順序回帰畳み込みネットワークを用いた非アルコール性脂肪肝炎における肝線維化診断
齋藤 僚介; 小泉 憲裕; 西山 悠; 今泉 飛翔; 草原 健太; 矢ヶ崎 詞穂; 小川 眞広; 松本 直樹
日本コンピュータ外科学会誌, (一社)日本コンピュータ外科学会, 22, 4, 281-281, Nov. 2020
Japanese

深層学習を用いた超音波プローブの位置推定における光学式データの活用の検討
矢ヶ崎 詞穂; 小泉 憲裕; 西山 悠; 近藤 亮祐; 草原 健太; 五十嵐 立樹; 齋藤 僚介; 沼田 和司; 小川 眞広; 松本 直樹
日本コンピュータ外科学会誌, (一社)日本コンピュータ外科学会, 22, 4, 303-303, Nov. 2020
Japanese

人工知能を用いた非アルコール性脂肪肝炎(NASH)の超音波診断システムの開発
齋藤 僚介; 小泉 憲裕; 西山 悠; 今泉 飛翔; 草原 健太; 矢ヶ崎 詩穂; 小川 眞広; 松本 直樹
超音波医学, (公社)日本超音波医学会, 47, Suppl., S541-S541, Nov. 2020
Japanese

深層学習を用いた超音波画像上の肝血管腫と血管の分類手法
草原 健太; 小泉 憲裕; 今泉 飛翔; 西山 悠; 齋藤 僚介; 矢ヶ崎 詞穂; 松本 直樹; 小川 眞広
超音波医学, (公社)日本超音波医学会, 47, Suppl., S542-S542, Nov. 2020
Japanese

超音波ガイド下ラジオ波焼灼療法支援システムにおける腫瘍追従手法に関する研究
矢ヶ崎 詞穂; 小泉 憲裕; 西山 悠; 近藤 亮祐; 今泉 飛翔; 草原 健太; 齋藤 僚介; 沼田 和司; 小川 眞広; 松本 直樹
超音波医学, (公社)日本超音波医学会, 47, Suppl., S542-S542, Nov. 2020
Japanese

超音波医工学を核とする医工融合人材の養成
小泉 憲裕; 西山 悠; 月原 弘之; 宮嵜 英世; 小路 直; 沼田 和司; 松本 直樹; 小川 眞広
超音波医学, (公社)日本超音波医学会, 47, Suppl., S565-S565, Nov. 2020
Japanese

Model-based kernel sum rule: kernel Bayesian inference with probabilistic models
Yu Nishiyama; Motonobu Kanagawa; Arthur Gretton; Kenji Fukumizu
Machine Learning, 109, 5, 939-972, 01 May 2020, Peer-reviwed
Scientific journal, English

Prediction of future gastric cancer risk using a machine learning algorithm and comprehensive medical check-up data: A case-control study
Junichi Taninaga; Yu Nishiyama; Kazutoshi Fujibayashi; Toshiaki Gunji; Noriko Sasabe; Kimiko Iijima; Toshio Naito
Scientific Reports, 9, 1, 12384-12384, 01 Dec. 2019, Peer-reviwed, True
Scientific journal, English

Automatic Diagnosis by Compact Portable Ultrasound Robot: State Estimation of Internal Organs with Steady-State Kalman Filter
Yudai Sasaki; Fumio Eura; Kento Kobayashi; Ryosuke Kondo; Kyohei Tomita; Yu Nishiyama; Hiroyuki Tsukihara; Naoki Matsumoto; Norihiro Koizumi
2019 IEEE Healthcare Innovations and Point of Care Technologies, HI-POCT 2019, 29-32, Nov. 2019
International conference proceedings, English

Development of bed-type ultrasound diagnosis and therapeutic robot
Kento Kobayashi; Yudai Sasaki; Fumio Eura; Ryosuke Kondo; Kyohei Tomita; Takahiro Kobayashi; Yusuke Watanabe; Akihide Otsuka; Hiroyuki Tsukihara; Naoki Matsumoto; Kazushi Numata; Hidetoshi Nagaoka; Toshiyuki Iwai; Hideyuki Iijima; Yu Nishiyama; Norihiro Koizumi
2019 IEEE International Conference on Cyborg and Bionic Systems, CBS 2019, -, -, 171-176, 18 Sep. 2019, Peer-reviwed
International conference proceedings, English

Deep learning based 3-dimensional liver motion estimation using 2-dimensional ultrasound images
Tsubasa Imaizumi; Ryosuke Kondo; Kenta Kusahara; Yu Nishiyama; Hiroyuki Tsukihara; Naoki Matsumoto; Kazushi Numata; Norihiro Koizumi
2019 IEEE International Conference on Cyborg and Bionic Systems, CBS 2019, -, -, 184-190, 18 Sep. 2019, Peer-reviwed
International conference proceedings, English

Lesion tracking method using CNN for non-invasive ultrasound theranostic system
Riki Igarashi; Kyohei Tomita; Yu Nishiyama; Yuka Shigenari; Norihiro Koizumi
2019 IEEE International Conference on Cyborg and Bionic Systems, CBS 2019, -, -, 228-234, 18 Sep. 2019, Peer-reviwed
International conference proceedings, English

Development of compact portable ultrasound robot for home healthcare
Yudai Sasaki; Fumio Eura; Kento Kobayashi; Ryosuke Kondo; Kyohei Tomita; Yu Nishiyama; Hiroyuki Tsukihara; Naoki Matumoto; Norihiro Koizumi
JOURNAL OF ENGINEERING-JOE, 14, 495-499, Feb. 2019, Peer-reviwed
Scientific journal, English

Method of medical digitalization（me-digit）and its effects and social impacts
Norihiro Koizumi; Yu Nishiyama; Fumio Eura; Tsubasa Imaizumi; Akihide Otsuka; Yudai Sasaki; Yuka Shigenari; Riki Igarashi; Kenta Kusahara; Kento Kobayashi; Hiroyuki Tsukihara; Naoki Matsumoto; Masahiro Ogawa; Sunao Shoji
Seimitsu Kogaku Kaishi/Journal of the Japan Society for Precision Engineering, 85, 9, 749-752, 2019
Scientific journal

Development of Compact Portable Ultrasound Robot for Home Healthcare
Yudai Sasaki; Fumio Eura; Kento Kobayashi; Ryosuke Kondo; Kyohei Tomita; Yu Nishiyama; Hiroyuki Tsukihara; Naoki Matumoto; Norihiro Koizumi
Proc. of 14th Annual Asian Conference on Computer Aided Surgery (ACCAS2018), 31-36, 16 Nov. 2018, Peer-reviwed
International conference proceedings, English

Prediction of glucose metabolism disorder risk using a machine learning algorithm: Pilot study
Katsutoshi Maeta; Yu Nishiyama; Kazutoshi Fujibayashi; Toshiaki Gunji; Noriko Sasabe; Kimiko Iijima; Toshio Naito
JMIR Diabetes, {JMIR} Publications Inc., 3, 4, e10212-?, 01 Oct. 2018, Peer-reviwed, True
Scientific journal, English

Automatic fascia extraction and classification for measurement of muscle layer thickness
Tsubasa Imaizumi; Norihiro Koizumi; Ryosuke Kondo; Yu Nishiyama; Naoki Matsumoto
2018 15th International Conference on Ubiquitous Robots, UR 2018, 493-496, 20 Aug. 2018, Peer-reviwed
International conference proceedings, English

Method for Extracting Acoustic Shadows to Construct an Organ Composite Model in Ultrasound Images
Akihide Otsuka; Norihiro Koizumi; Izumu Hosoi; Hiroyuki Tsukihara; Yu Nishiyama
2018 15th International Conference on Ubiquitous Robots, UR 2018, 719-722, 20 Aug. 2018, Peer-reviwed
International conference proceedings, English

Out-of-Plane Motion Detection System Using Convolutional Neural Network for US-guided Radiofrequency Ablation Therapy
Ryosuke Kondo; Norihiro Koizumi; Yu Nishiyama; Naoki Matsumoto; Kazushi Numata
2018 15th International Conference on Ubiquitous Robots, UR 2018, 729-731, 20 Aug. 2018, Peer-reviwed
International conference proceedings, English

Method for Shape Extraction and Modeling of Prostate Contours by Using Sperellipses
Yuka Shigenari; Norihiro Koizumi; Yu Nishiyama
Proc. of 15th International Conference on Ubiquitous Robots (UR), The Japan Society of Mechanical Engineers, 2018, ?-?, 26 Jun. 2018, Peer-reviwed, In this report we propose a novel method for shape extraction and modeling of MRI prostate contours. Transperineal targeted biopsy with real-time fusion image of multiparametric magnetic resonance image and transrectal ultrasound image have a problem that the diagnostic ability is uneven in accordance with the skill of the medical professionals. To cope with these problems, we propose a method, which is not affected by diagnostic ability, to extract prostate contour utilizing deformable superellipse models. The characteristic of oue method is to implement the region division function for the deformable superellips models to enhance the expressiveness to handle the left and right asymmetry of the prostate contours. Experimental results show our proposing method is effectiveness.
International conference proceedings, English

Deep Learningを用いたオペレータへの予測画像提示による遅延補償
松尾開; 西山悠; 小木曽公尚; 稲垣哲哉; 浜本研一
第５回制御部門マルチシンポジウム (MSCS2018), Fr73-1, ?, ?-?, 09 Mar. 2018
Symposium, Japanese

Short-Term Wind-Speed Forecasting Using Kernel Spectral Hidden Markov Models.
Shunsuke Tsuzuki; Yu Nishiyama
CoRR, abs/1811.06210, 2018
Scientific journal

医療・バイオは新しいデジタルだ－医デジ化による超高精度な超音波診断・治療の実現－
小泉憲裕; 徐 俊浩; 李 得熙; 栢菅 篤; 近藤亮祐; 冨田恭平; 細井泉澄; 西山 悠; 月原 弘之; 宮嵜 英世; 福田浩之; 沼田 和司; 葭仲 潔; 東 隆; 杉田直彦; 本間之夫; 松本洋一郎; 光石 衛
Jpn J Med Ultrasonics, The Japan Society of Ultrasonics in Medicine, 45, 2, 173-182, 2018, Peer-reviwed, The expectation that intersections of various science and engineering technologies such as mathematics, information, control, artificial intelligence, and robot technology with medicine and biology have enormous potential is rapidly growing day by day. Bill Gates said, “If I were a student, I would learn biology,” and Nicolas Negroponte said, “Bio is new digitals.” It is a very clear phrase that predicts that biology will be reconstructed by the fusion of bio and IT technology. Here, the meaning of bio is intended to cover a wide range of bio including biotechnology. In this paper, we review medical digitization (Me-DigIT) in the field of medical ultrasound and core technologies for Me-DigIT.
Scientific journal, Japanese

Characteristic extraction for model parameters of McKibben pneumatic artificial muscles
T. Ishikawa; Y. Nishiyama; K. Kogiso
SICE Journal of Control, Measurement, and System Integration, 11, 4, 357-364, 2018, Peer-reviwed
Scientific journal, English

エピゲノム解析用検体自動調製システムを用いた生命科学実験の自動化
三木 陽太; 大橋 一晶; 加藤 太輔; 土屋 正年; 小泉 憲裕; 小木曽 公尚; 西山 悠; 戸澤 英人; 田口 明糸; 小林 美佳; 和田 洋一郎; 井原 茂男
第60回自動制御連合講演会, The Japan Joint Automatic Control Conference, SuA1-1, ?, ?-?, 12 Nov. 2017
Symposium, Japanese

医療・バイオ分野のデジタル化を加速する医学デジ化コア技術
小泉 憲裕; 西山 悠; 小木曽 公尚; 和田 洋一郎
第60回自動制御連合講演会, The Japan Joint Automatic Control Conference, SuA1-2, ?, ?-?, 12 Nov. 2017
Symposium, Japanese

患部追従超音波プローブを用いた次世代型生体モニタリング装置の開発プロジェクト
江浦 史生; 相澤 理佳; 西山 悠; 近藤 亮祐; 冨田 恭平; 小木曽 公尚; 小泉 憲裕
第60回自動制御連合講演会, The Japan Joint Automatic Control Conference, SuA1-4, ?, ?-?, 12 Nov. 2017
Symposium, Japanese

Parameter extraction for identifying product type of mckibben pneumatic artificial muscles
Takahiro Ishikawa; Yu Nishiyama; Kiminao Kogiso
1st Annual IEEE Conference on Control Technology and Applications, CCTA 2017, 2017-January, 1935-1940, 06 Oct. 2017, Peer-reviwed
International conference proceedings, English

Efficient PSO-based algorithm for parameter estimation of McKibben PAM model
Atsushi Okabe; Takahiro Ishikawa; Kiminao Kogiso; Yu Nishiyama
1st Annual IEEE Conference on Control Technology and Applications, CCTA 2017, 2017-January, 1414-1419, 06 Oct. 2017, Peer-reviwed
International conference proceedings, English

A study for tracking focal lesions in non-invasive ultrasound theragnostic system
Kyohei Tomita; Norihiro Koizumi; Atsushi Kayasuga; Yu Nishiyama; Hiroyuki Tsukihara; Hideyo Miyazaki; Kiyoshi Yoshinaka; Mamoru Mitsuishi
2017 14th International Conference on Ubiquitous Robots and Ambient Intelligence, URAI 2017, 589-591, 25 Jul. 2017, Peer-reviwed
International conference proceedings, English

An automatic templates selection method for ultrasound guided tumor tracking
Ryosuke Kondo; Norihiro Koizumi; Kyohei Tomita; Yu Nishiyama; Hidenori Sakanashi; Hiroyuki Fukuda; Hiroyuki Tsukihara; Kazushi Numata; Mamoru Mitusishi; Yoichiro Matsumoto
2017 14th International Conference on Ubiquitous Robots and Ambient Intelligence, URAI 2017, 587-588, 25 Jul. 2017, Peer-reviwed
International conference proceedings, English

Development of portable ultrasound guided physiological motion compensation device
Fumio Eura; Rika Aizawa; Ryousuke Kondo; Kyohei Tomita; Yu Nishiyama; Norihiro Koizumi
2017 IEEE International Conference on Cyborg and Bionic Systems, CBS 2017, 2018-January, 243-247, 02 Jul. 2017, Peer-reviwed
International conference proceedings, English

肝線維化の定量化を目的とするテクスチャ解析を用いたアーチファクトの分類手法
源田 達也; 小泉 憲裕; 大塚 研秀; 近藤 亮祐; 冨田 恭平; 西山 悠; 坂無 英徳; 熊川 まり子; 松本 直樹; 小川 眞広
超音波医学, (公社)日本超音波医学会, 44, Suppl., S457-S457, Apr. 2017
Japanese

超音波による人体検出機能を持つ移動ロボットの開発
苗村智行; 山本健司; 西山悠; 植野真臣
平成29年電気学会全国大会, 0, 0, 0-0, 17 Mar. 2017
Symposium, Japanese

超音波画像における音響シャドウを除去した臓器合成モデルの構築法
細井 泉澄; 小泉 憲裕; 栢菅 篤; 冨田 恭平; 西山 悠; 月原 弘之; 福田 浩之; 葭中 潔; 斎藤 季; 宮崎 英世; 杉田 直彦; 沼田 和司; 本間 之夫; 松本 洋一郎; 光石 衛
精密工学会学術講演会講演論文集, 公益社団法人 精密工学会, 2017, 233-234, 2017, 肋骨存在下の超音波診断においては、肋骨等によって生じる音響シャドウのために、医療専門家が本当に確認したい方向からの診断画像を得ることができないという大きな問題がある。上記を踏まえて本報では、肋骨存在下の腎臓の超音波診断を対象に、臓器の輪郭情報を手がかりに複数枚の臓器画像を合成することで、音響シャドウの影響を除去した超音波診断画像を再構築する手法を世界で初めて開発したので、これについて報告する。
Japanese

A robust and precise focal lesion servo method for non-invasive ultrasound theragnostic system
TOMITA Kyohei; KOIZUMI Norihiro; KAYASUGA Atsushi; NISHIYAMA Yu; TSUKIHARA Hiroyuki; MIYAZAKI Hideyo; YOSHINAKA Kiyoshi; MITSUISHI Mamoru
The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), The Japan Society of Mechanical Engineers, 2017, 1P1-I05, 2017,

In recent years, HIFU (High Intensity Focused Ultrasound) therapy, which is one of the non-invasive therapies utilizing focused ultrasound, have attracted a great attention as a novel treatment method for a focal lesion, such as a tumor and a stone. However, the focal lesion moves in accordance with respiration, which may cause the damage for the surrounding normal tissues. To cope with this problem, we have developed a noninvasive ultrasound theragnostic system (NIUTS). In this report, we proposed a novel tracking method, which is implemented in NIUTS, based on "Partial Active Shape Model" to enhance the servo performance for the focal lesion. Experimental results shows the effectiveness of the proposed servo method concerning the precision and robustness for a kidney phantom.

Japanese

An ultrasound guided tracking method for a tumor utilizing HLAC based dynamic template matching
KONDO Ryosuke; KOIZUMI Norihiro; TOMITA Kyohei; NISHIYAMA Yu; FUKUDA Hiroyuki; TSUKIHARA Hiroyuki; NUMATA Kazushi; MATSUMOTO Yoichiro; MITSUISHI Mamoru
The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), The Japan Society of Mechanical Engineers, 2017, 1A1-L04, 2017,

In this report we propose a novel robust tumor tracking method for ultrasound guided RFA treatments. RFA (radiofrequency ablation) treatments have a serious problem to hide a tumor to be treated due to the hyperechoic region, which is generated by vapors during RFA. Moreover, organ deformations seriously deteriorate the tracking performance. To cope with these problems, we propose a novel method to track a tumor in ultrasound diagnostic image. Templates, which is used to track the tumor, are generated dynamically by HLAC(Higher order Local Auto Correlation) -based subspace method. Our method achieves stable tracking by selecting templates automatically based on the texture features of ultrasound diagnostic images, while the conventional method is unstable due to the variation to select templates manually. Experimental results show the effectiveness of our proposing motion tracking method concerning the robustness and accuracy.

Japanese

Acoustic-shadow-free organ composition image medical ultradound
HOSOI Izumu; Koizumi Norihiro; TOMITA Kyohei; NISHIYAMA You; TSUKIHARA Hiroyuki; MIYAZAKI Hideyo; YOSHINAKA Kiyoshi; MITSUISHI Mamoru
The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), The Japan Society of Mechanical Engineers, 2017, 1A1-J05, 2017,

Acoustic shadows, which are generated by rib bones, make it difficult to obtain proper diagnostic images from the proper viewpoints for medical professionals. To cope with this, we combined separate parts of organ images together by utilizing the contours of the kidney. The contours are extracted by the algorithms which are Snake and Partial Active Shape Model. And we constructed the organ composition model that removes the acoustic shadow. It is confirmed that the composition model could be generated properly by the proposed method.

Japanese

McKibben型空気圧ゴム人工筋モデルの特徴的なパラメータの抽出
石川貴大; 西山悠; 小木曽公尚
第59回自動制御連合講演会, 0, 0, 0-0, 11 Nov. 2016
Symposium, Japanese

Co-creating ルーブリックの自己制御学習および自己評価力への影響分析
山本美紀; 宇都雅輝; 西山悠; 川野秀一; 植野真臣
日本テスト学会 第14回大会, 86-87, 09 Sep. 2016
Symposium, Japanese

Characteristic kernels and infinitely divisible distributions
Yu Nishiyama; Kenji Fukumizu
Journal of Machine Learning Research, 17, 180, 180-28, 01 Sep. 2016, Peer-reviwed
Scientific journal, English

McKibben型空気圧ゴム人工筋の製品種別に関する識別器の構成
石川貴大; 岡部篤; 西山悠; 小木曽公尚
第3回計測自動制御学会 制御部門マルチシンポジウム(MSCS2016), 0, 0, 0-0, 07 Mar. 2016
Symposium, Japanese

Filtering with state-observation examples via kernel Monte Carlo filter
Motonobu Kanagawa; Yu Nishiyama; Arthur Gretton; Kenji Fukumizu
Neural Computation, 28, 2, 382-444, 01 Feb. 2016, Peer-reviwed, True
Scientific journal, English

The nonparametric kernel bayes smoother
Yu Nishiyama; Amir Hossein Afsharinejad; Shunsuke Naruse; Byron Boots; Le Song
Proceedings of the 19th International Conference on Artificial Intelligence and Statistics, AISTATS 2016, 51, 00, 547-555, 09 May 2016, Peer-reviwed
International conference proceedings, English

Bayes factor-based learning Bayesian networks
Kazuki Natori; Masaki Uto; Yu Nishiyama; Shuichi Kawano; Maomi Ueno
2nd Workshop on Advanced Methodologies for Bayesian Networks, 15-31, 16 Nov. 2015, Peer-reviwed
International conference proceedings, English

Constraint-based learning Bayesian networks using Bayes factor
Kazuki Natori; Masaki Uto; Yu Nishiyama; Shuichi Kawano; Maomi Ueno
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Springer, 9505, 15-31, 16 Nov. 2015, Peer-reviwed
International conference proceedings, English

Monte Carlo filtering using kernel embedding of distributions
Motonobu Kanagawa; Yu Nishiyama; Arthur Gretton; Kenji Fukumizu
Proceedings of the National Conference on Artificial Intelligence, 3, 1897-1903, 2014, Peer-reviwed
International conference proceedings, English

Hilbert Space Embeddings of POMDPs
Yu Nishiyama; Abdeslam Boularias; Arthur Gretton; Kenji Fukumizu
CoRR, abs/1210.4887, 2012
Scientific journal

A family of CCCP algorithms which minimize the TRW free energy
Yu Nishiyama; Xingyao Ye; Alan L. Yuille
New Generation Computing, 30, 1, 3-16, Jan. 2012, Peer-reviwed
International conference proceedings, English

Hilbert space embeddings of pomdps
Yu Nishiyama; Abdeslam Boularias; Arthur Gretton; Kenji Fukumizu
Uncertainty in Artificial Intelligence - Proceedings of the 28th Conference, UAI 2012, AUAI Press, abs/1210.4887, 644-653, 2012, Peer-reviwed
International conference proceedings, Japanese

Accuracy of Loopy belief propagation in Gaussian models
Yu Nishiyama; Sumio Watanabe
Neural Networks, 22, 4, 385-394, May 2009, Peer-reviwed, True
Scientific journal, English

On the minima of bethe free energy in gaussian distributions
Yu Nishiyama; Sumio Watanabe
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 5097 LNAI, 1075-1086, 2008, Peer-reviwed
International conference proceedings, English

Generalization of concave and convex decomposition in Kikuchi free energy
Yu Nishiyama; Sumio Watanabe
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 5163 LNCS, PART 1, 51-60, 2008, Peer-reviwed
International conference proceedings, English

Stochastic complexity of complete bipartite graph-type Boltzmann machines in mean field approximation
Yu Nishiyama; Sumio Watanabe
Electronics and Communications in Japan, Part III: Fundamental Electronic Science (English translation of Denshi Tsushin Gakkai Ronbunshi), 90, 9, 1-9, Sep. 2007
Scientific journal, English

Theoretical analysis of accuracy of gaussian belief propagation
Yu Nishiyama; Sumio Watanabe
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 4668 LNCS, PART 1, 29-38, 2007, Peer-reviwed
International conference proceedings, English

Stochastic Complexity of Complete Bipartite Graph-Type Boltzmann Machines in Mean Field Approximation
NISHIYAMA Yu; WATANABE Sumio
The Transactions of the Institute of Electronics, Information and Communication Engineers. A, The Institute of Electronics, Information and Communication Engineers, 89, 8, 671-678, 01 Aug. 2006, Peer-reviwed, ベイズ事後分布を比較的少ない計算量で実現するための近似手法として変分ベイズ法が提案され,実問題への有効性が確認されている.変分ベイズ法は統計物理学において分配関数の計算に用いられる平均場近似を一般化した方法であり,近年,その近似精度などの数理的な性質についても研究が行われている.本論文では,完全2部グラフ型ボルツマンマシンに平均場近似を適用した場合の確率的複雑さの漸近形について考察し,その漸近形を理論的に導出する.また,その結果に基づいて,ベイズ事後分布と平均場近似による事後分布との相違について定量的な考察を行う.
Scientific journal, Japanese

Asymptotic behavior of stochastic complexity of complete bipartite graph-type boltzmann machines
Yu Nishiyama; Sumio Watanabe
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 4232 LNCS, 417-426, 2006, Peer-reviwed
International conference proceedings, English

Development of a system to imitation the motion trajectory of a probe using Visual SLAM for an automated ultrasound diagnostic robot
Monma Sho; Koizumi Norihiro; Nishiyama Yu; Ishikawa Tomohiro; Jiayi Zhou; Watanabe Yusuke; Fujibayashi Takumi; Matsuyama Momoko; Yamada Miyu; Tsumura Ryosuke; Yoshinaka Kiyoshi; Matsumoto Naoki; Ogawa Masahiro; Tsukihara Hiroyuki; Numata Kazushi
The purpose of this study is to estimate the motion trajectory of an ultrasound probe using Visual SLAM technology，and to reproduce a doctor's probe scanning using a bed-type robotic ultrasound diagnosis system (RUDS)． We investigated a method to mimic the motion trajectory of the RUDS probe by using SLAM technology．The proposed method was able to perform the imitation motion，but the error became larger as the distance from the origin increased．, The Japan Society of Mechanical Engineers, Jun. 2022, The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), 2022, 1P1-M02, Japanese, 2424-3124

Proposal of an Automatic Probe Manipulation Model Considering Acoustic Shadow in Ultrasound Diagnosis
MATSUYAMA Momoko; KOIZUMI Norihiro; NISHIYAMA Yu; WATANABE Yusuke; ZHOU Jiayi; YAGASAKI Shiho; FUJIBAYASHI Takumi; YAMADA Miyu; ISHIKAWA Tomohiro; TSUMURA Ryosuke; YOSHINAKA Kiyoshi; MATSUMOTO Naoki; TSUKIHARA Hiroyuki; NUMATA Kazushi
In ultrasound therapy, a clear ultrasound image is necessary to determine the exact irradiation position. However, there is a concern that the accuracy of irradiation may be degraded due to the black noise caused by the reflection of sound waves on hard tissues such as ribs and stones. In this study, we aim to automate ultrasound probe manipulation to support monitoring of ultrasound diagnosis. The acoustic shadow and the target organ in the ultrasound image are detected by deep learning, and the control model avoids overlapping imaging in real time based on the overlapping area information of the two. This makes it possible to monitor the treatment target without any acoustic shadows., The Japan Society of Mechanical Engineers, Jun. 2022, The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), 2022, 1P1-M11, Japanese, 2424-3124

Detection of Organs and Identification of Missing Diagnostic Images Using Deep Learning for Assessment of Diagnostic Image Adequacy
KATSURAGI Arashi; KOIZUMI Norihiro; NISHIYAMA Yuu; ZHOU Jiayi; WATANABE Yusuke; FUJIBAYASHI Takumi; MATSUYAMA Momoko; YAMADA Miyu; YOSHINAKA Kiyoshi; TSUMURA Ryosuke; TSUKIHARA Hiroyuki; NUMATA Kazushi; MATSUMOTO Naoki; MASUZAKI Ryouta; OGAWA Masahiro
The purpose of this study was to evaluate the appropriateness of diagnostic images for automated ultrasound operations. Therefore, two experiments were conducted. The first is to detect the target organ using deep learning. The second is to identify missing parts in the diagnostic images. In the first experiment of organ detection, the IoU and Dice coefficient were 0.947 and 0.972, respectively, indicating high accuracy.In the second experiment to identify the missing parts of the image, the percentage of correct answers for the missing parts on the right side of m was 75.3%, while the percentage of correct answers for the missing parts on the left side was 99.1%., The Japan Society of Mechanical Engineers, Jun. 2022, The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), 2022, 1P1-M12, Japanese, 2424-3124

Construction of a 3D Model of Kidney Using Visual SLAM and Deep Learning with Robotic Ultrasound
ISHIKAWA Tomohiro; KOIZUMI Norihiro; NISHIYAMA Yu; ZHOU Jiayi; WATANABE Yusuke; FUJIBAYASHI Takumi; MATSUYAMA Momoko; YAMADA Miyu; TSUKIHARA Hiroyuki; NUMATA Kazushi; YOSHINAKA Kiyoshi; TSUMURA Ryosuke
Ultrasonography is less invasive and safer than MRI or CT. However, image acquisition is dependent on the skill of the probe operator, and it is difficult for an untrained examiner to understand the three-dimensional structure of the organ. For the above reason, we use a robot to acquire ultrasound images and estimate the position of the probe using Visual SLAM. The obtained images are segmented and combined with position information to construct a three-dimensional model of the organ. In this study, the right kidney of the phantom was used as the target, and the results of position estimation and segmentation accuracy were high. The model of the right kidney was constructed based on these results., The Japan Society of Mechanical Engineers, Jun. 2022, The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), 2022, 1P1-L04, Japanese, 2424-3124

Development of an Intraoperative Cancer Localization Navigation System to Support Complete Resection of Prostate Cancer
MUKASA Anju; KOIZUMI Norihiro; NISHIYAMA Yu; ONODERA Yusuke; MATSUYAMA Momoko; FUJIBAYASHI Takumi; SHOJI Sunao
In this study, we investigated the segmentation and registration methods to develop an intraoperative cancer localization navigation system to support complete resection of prostate cancer. segmentation using YOLACT++ was very accurate in the prostate. YOLACT++ was able to segment the prostate with very high accuracy, and it was also able to obtain sufficient accuracy for the outline of the tumor near the capsule of prostate. Therefore, we evaluated the effectiveness of YOLACT++ for cancer localization prediction. For registration, we compared the accuracy of affine transformation and projection transformation, and evaluated that affine transformation was effective. In the future, we will examine the accuracy of nonlinear registration methods such as the B-spline method and registration methods between different modalities., The Japan Society of Mechanical Engineers, Jun. 2022, The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), 2022, 1P1-L12, Japanese, 2424-3124

A contact state adjustment method to enhance organ motion compensation performance for a bed-type ultrasound diagnostic and therapeutic robot
KOBAYASHI Kento; SASAKI Yudai; KOBAYASHI Takahiro; WATANABE Yusuke; ZHOU Jiayi; NISHIYAMA Yu; KOIZUMI Norihiro; TSUKIHARA Hiroyuki; NUMATA Kazushi; IIJIMA Hideyuki; Iwai Toshiyuki; Nagaoka Hidetoshi
It is difficult to perform proper ultrasound diagnosis when the respiratory organs are moving, so it is necessary for the patient to stop breathing. However, stopping breathing puts a burden on the patient, so it is necessary to acquire a still image while breathing, that is, even when the organ is moving. We developed a bed-type ultrasound robot and used template matching for tracking and load cell for measuring the contact force. The subject of the tracking experiment was a phantom that imitated the abdominal organ, and the contact force was measured using the phantom. Tracking accuracy was higher at a bed speed of 1.4 cm/s on the back when the contact was sufficient on the abdomen. As for the contact force, a value of 3.0 to 4.9 N was measured on the back., The Japan Society of Mechanical Engineers, 2020, The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), 2020, 2A1-E03, Japanese, 2424-3124, 130007943902

Development of vertical axis rotation mechanism of bed type ultrasonic robot and estimation method of posture of moving organ
KOBAYASHI Takahiro; SASAKI Yudai; KOBAYASHI Kento; WATANABE Yusuke; ZHOU Kai; NISHIYAMA Yu; KOIZUMI Norihiro; TSUKIHARA Hiroyuki; NUMATA Kazushi; IIJIMA Hideyuki; IWAI Toshiyuki; NAGAOKA Hidetoshi
In this research, we developed a mechanism that has vertical axis rotation in the tip of a bed-type ultrasound robot that can follow three-dimensional movement of abdominal organs with higher accuracy, which is impossible with a conventional bed-type ultrasound robot. As a preparation stage for position and orientation tracking, a new method for estimating the organ angle from acquired ultrasound images alone is newly proposed. From the experimental results, an appropriate angle could be estimated from the acquired data. As future work, it is required to perform an experiment to follow the organ in the vertical axis rotation with the developed mechanism and the proposed algorithm., The Japan Society of Mechanical Engineers, 2020, The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), 2020, 2A1-E12, Japanese, 2424-3124, 130007943854

Study on automatic acquisition of diagnostic images by ultrasound diagnostic robot
Watanabe Yusuke; Sasaki Yudai; Kobayashi Kento; Kobayashi Takahiro; Jiayi Zhou; Nishiyama Yu; Koizumi Norihiro; Tsukihara Hiroyuki; Numata Kazushi; Iijima Hideyuki; Iwai Toshiyuki; Nagaoka Hidetoshi
The purpose of this study is to develop a novel algorithm that can automatically acquire an ultrasound image using the newly developed bed-type ultrasound diagnostic robot. We confirmed the effectiveness of our novel proposed method by comparing tracking accuracy with the conventional template matching method and simulated a basic algorithm for controlling the robot. As a result, it was confirmed that the accuracy for detecting the target is dramatically enhanced so as to obtain the target ultrasound image successfully., The Japan Society of Mechanical Engineers, 2020, The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), 2020, 2A1-E14, Japanese, 2424-3124, 130007943853

An Automated Ultrasound Diagnosis System for Liver Fibrosis in Non-Alcoholic Steatohepatitis Using Deep Learning
SAITO Ryosuke; KOIZUMI Norihiro; NISHIYAMA Yu; IMAIZUMI Tsubasa; KUSAHARA Kenta; YAGASAKI Shiho; OGAWA Masahiro; MATSUMOTO Naoki
Liver fibrosis is important information for diagnosing the prognosis of fatty liver. Diagnosis of the degree of fibrosis by ultrasound is non-invasive and cost-effective. However, it is difficult to evaluate the effect of fat on interpretation and mild fibrosis. In this report, we propose a novel method utilizing deep learning to improve the accuracy and automation of ultrasound diagnosis of liver fibrosis for NASH. This is a novel system that extracts the parenchyma of liver by U-Net, and then performs classification using the network that considers the order of the fibrosis level. The experimental results showed that the extraction of parenchymal liver achieved a Dice coefficient of 0.929, demonstrating the effectiveness of the method using U-Net. As for the classification, the accuracy rate was improved to 0.639 than that of the conventional method., The Japan Society of Mechanical Engineers, 2020, The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), 2020, 2A1-E15, Japanese, 2424-3124, 130007943852

Estimation of Organ State utilizing Deep Learning Tiny-YOLOv3 and Fixed Lag Smoothing Motion Filter
ZHOU Jiayi; SASAKI Yudai; KOBAYASHI Kento; EURA Fumio; NISHIYAMA Yu; TSUKIHARA Hiroyuki; MATSUMOTO Naoki; KOIZUMI Norihiro
In this research, we proposed a robotic motion control framework that can detect kidney in real-time by utilizing deep learning, and evaluate the accuracy of automatically acquiring and maintaining ultrasound diagnostic images of kidney. Furthermore, we performed object detection experiments using a model that was generated by a kidney phantom, estimated the state of renal phantom motion.

The novelty of our method is the framework of the combination of the deep learning tiny-YOLOv3 model and filtering considering the influence of speckle noise in an ultrasound image. In our method, the filtering is determined to incorporate the center position of where the object is detected, and two types of filters are adopted., The Japan Society of Mechanical Engineers, 2020, The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), 2020, 2A1-E02, Japanese, 2424-3124, 130007943896

Comparison and Validation of Fast Organ Detection Methods on Ultrasound Images Using CNN
IGARASHI Riki; TOMITA Kyohei; NISHIYAMA Yu; KOIZUMI Norihiro
We are developing a system to treat kidney, liver and other organs, as well as stones and cancer in these organs using high-power focused ultrasound (HIFU) while tracking lesions that move by breathing and body movements. The system estimates organ movement by analyzing ultrasound images obtained from the probe and compensates for this movement with robotic control. In recent years, various medical image analysis methods using deep learning have been proposed and studied, but they have not been fully explored as methods to detect specific organs using robotic systems. In this paper, we compared and validated the performance of Faster R-CNN, which is commonly used for object detection, and the proposed methods, Regression Network (RegNet) and Segmentation In Regression Network (SegInRegNet), the proposed method based on the problems of Faster R-CNN, in a kidney detection task. And then, we show that 1) there are some problems with Faster R-CNN as a kidney detection method operating on a robotic system and that 2) proposed method performs better than Faster R-CNN in terms of detection speed and accuracy., The Society of Instrument and Control Engineers, 2020, Transactions of the Society of Instrument and Control Engineers, 56, 12, 560-569, Japanese, 0453-4654, 1883-8189, 130007958053, AN00072392

Matching axial images of magnetic resonance imaging and transrectal ultrasound based on deep learning
Igarashi Riki; Koizumi Norihiro; Nishiyama Yu; Tomita Kyohei; Shigenari Yuka; Shoji Sunao
This paper examines the feasibility of automated alignment in prostate targeted biopsy by comparing the prostate contour between different modalities. The prostate targeted biopsy that is attracting attention in the treatment of prostate cancer largely depends on the doctor who operates surgery, so it can be expected to reduce the variation in the diagnostic performance by automation. In the proposed method, segmentation is performed using deep learning, and the same prostate cross section between different modalities is estimated from the similarity obtained by comparing prostate contours of different modalities obtained by segmentation. In this method it was possible to estimate close to expert judgment with accuracy of 69.4%. Furthermore, by considering the deformity of the prostate gland and calculating the similarity for each angle, we achieved an estimate close to the judgment of experts with higher accuracy of 83.3%., The Japan Society of Mechanical Engineers, 2019, The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec), 2019, 1A1-B09, Japanese, 2424-3124, 130007774097

医デジ化の手法・効果とこれがもたらす社会的インパクト
小泉憲裕; 西山悠; 江浦史生; 今泉飛翔; 大塚研秀; 佐々木雄大; 重成佑香; 五十嵐立樹; 草原健太; 小林賢人; 月原弘之; 松本直樹; 小川眞広; 小路直
The Japan Society for Precision Engineering, 2019, Journal of the Japan Society for Precision Engineering, 85, 9, 749-752, Japanese, Peer-reviwed, Introduction scientific journal, 0912-0289, 1882-675X, 130007703006, AN1003250X

Extraction of characteristic parameter of McKibben pneumatic artificial muscle model
Ishikawa T.; Nishiyama Y.; Kogiso K.
This study extracts characteristic model parameters for McKibben pneumatic artificial muscle (PAM) system by using support vector machine. Five types of PAMs are used to evaluate the resulting classifies based on several combination of the parameters in terms of a percentage of correctness. Additionally, the classifiers and discriminators obtained using the extracted parameters are shown., The Japan Joint Automatic Control Conference, 2016, Proceedings of the Japan Joint Automatic Control Conference, 59, 711-713, Japanese, 130005312580

カーネルベイズスムージングとカーネル平均Toolboxの作成
西山 悠
2015, 第25回日本神経回路学会全国大会予稿集, 1, 60-61, False

部分観測マルコフ決定過程ベルマン方程式のカーネル化—Kernel Bellman Equations in POMDPs—情報論的学習理論と機械学習
西山 悠; Abdeslam Boularias; Arthur Gretton
電子情報通信学会, Mar. 2012, 電子情報通信学会技術研究報告 = IEICE technical report : 信学技報, 111, 480, 35-42, Japanese, 0913-5685, 110009545973, AA1123312X

ポスター講演 CCCPを用いたTRW自由エネルギー最小化に基づく確率推論—Probabilistic inference by minimizing the TRW free energy using CCCP—情報論的学習理論と機械学習
西山 悠; Xingyao Ye; Alan L. Yuille
電子情報通信学会, Nov. 2010, 電子情報通信学会技術研究報告 = IEICE technical report : 信学技報, 110, 265, 51-58, Japanese, 0913-5685, 110008153961, AA1123312X

A CDMA Multiuser Demodulation Algorithm Based on NCCCP
NISHIYAMA Yu; TONOSAKI Yukinori; WATANABE Sumio
For CDMA demodulation problem in mobile communication technology, analyses of performance and developments of algorithms on the basis of statistical mechanical informatics have been done. While CDMA multiuser demodulation algorithm based on belief propagation(BP) shows its efficacy especially in large systems with respect to users and chips, that based on CCCP is known to provide better performance in small systems, which often appear in practical situation. However, CCCP has a drawback that it requires huge computational cost. Recently, we proposed a new CCCP(NCCCP) algorithm, which is an extension of CCCP algorithm for the Bethe free energy and show it can reduce expensive computational cost. In this report, we develop the CDMA multiuser demodulation algorithm based on NCCCP and compare the performance with that of conventional CCCP., The Institute of Electronics, Information and Communication Engineers, 13 Dec. 2008, IEICE technical report, 108, 372, 49-54, Japanese, 0913-5685, 110007123457, AN10091178

An Extension of CCCP Algorithm for Bethe Free Energy
NISHIYAMA Yu; WATANABE Sumio
Belief Propagation (BP) is an efficient algorithm for computing marginal probabilities of a high-dimensional probability distribution. The marginals computed by BP are equivalent to the extrema of Bethe free energy. Concave convex procedure (CCCP) has been studied for the optimization of Bethe free energy. In this paper, we extend the CCCP algorithm for Bethe free energy and present a new CCCP (NCCCP) algorithm. We practically apply NCCCP algorithm to multi-dimensional Gaussian distributions. As a result, NCCCP algorithm enables inner loop to converge even if all Lagrange multipliers in the loop are simultaneously updated. Moreover, we find that there exists an optimal point of the parameters introduced to NCCCP that can reduce the expensive computational cost., The Institute of Electronics, Information and Communication Engineers, 05 Mar. 2008, IEICE technical report, 107, 542, 85-90, Japanese, 0913-5685, 110006783205, AN10091178

25pPSB-4 An Extension of CCCP Algorithm for Kikuchi Free Energy
Nishiyama Yu; Watanabe Sumio
The Physical Society of Japan (JPS), 29 Feb. 2008, Meeting abstracts of the Physical Society of Japan, 63, 1, 331-331, Japanese, 1342-8349, 110007196251, AA11439205

Analytical and Numerical Solutions of Bethe Approximation in Normal Distributions
NISHIYAMA Yu; WATANABE Sumio
For the calculation of marginal distributions, which require the huge computational cost, the algorithm called belief propagation(BP) has shown the effectiveness. Loopy belief propagation(LBP), which is the BP applied to the distributions that have loops, is not guaranteed to converge in general and, if it converges, it computes approximate marginal probabilities. The LBP fixed-points after the convergence are also given by the extrema of the evaluation function called Bethe free energy. In this paper, we clarify the extrema of Bethe free energy in some particular cases of normal distributions, and after we give the convergence condition of LBP and the accuracy of LBP, we do the numerical experiments and compare the results with analytical solutions., The Institute of Electronics, Information and Communication Engineers, 15 Jan. 2008, IEICE technical report, 107, 413, 1-6, Japanese, 0913-5685, 110006623363, AN10091178

菊池自由エネルギーに対するCCCPアルゴリズムの拡張
西山悠
2008, DEX-SMI成果発表会, 2008, 10028997528

Theoretical Analysis of Accuracy of Belief Propagation in Gaussian Models
NISHIYAMA Yu; WATANABE Sumio
Belief propagation (BP) is and algorithm which can compute marginal probability distributions with a tractable computational cost. Loopy belief propagation (LBP) applied to the graphs containing loops is known to provide marginal distributions approximately if LBP converges. In this paper, we apply LBP to a multi-dimensional Gaussian distribution that has loops and analytically show how accurate LBP is in some cases. Specifically, we analytically show messages, approximate marginal densities, and the KL distance at fixed points of LBP when the graph corresponding to a Gaussian distribution has at most a single loop. Basides, for the graphs which have arbitrary structures, we derive the expansions of approximate marginal densities when covariances are small., The Institute of Electronics, Information and Communication Engineers, 14 May 2007, IEICE technical report, 107, 50, 23-28, English, 0913-5685, 110006291378, AN10091178

Asymptotic Behavior of Free Energy of General Boltzmann Machines in Mean Field Approximation
NISHIYAMA Yu; WATANABE Sumio
In the Bayesian learning, which generally requires huge computational costs, the algorithms based on the mean field approximation have shown us the effectiveness in the practical information systems. Recently, the generalization error or free energy in the mean field approximation has been theoretically studied. The theoretical results enable us to know the accuracy of the approximation and contribute to the foundation of a model selection in statistical singular machines. In this paper, we show that the upper bounds of the asymptotic free energies are theoretically obtained by counting the number of non-0 eigenvalues of Fisher information matrices and derive the upper bound in the learning model of general Boltzmann machines., The Institute of Electronics, Information and Communication Engineers, 07 Jul. 2006, IEICE technical report, 106, 163, 1-6, Japanese, 0913-5685, 110004809698, AN10091178

Stochastic Complexity of Complete Bibpartite Graph-type Boltzmann Machines in Mean Field Approximation
NISHIYAMA Yu; WATANABE Sumio
In the learning of singular learning machines, the superiority of Bayesian learning is shown. However, it requires huge computational costs to realize the Bayesian a posteriori distribution. To overcome this problem, the mean field approximation, which is originally known in statistical physics, is used in the practical information systems. Recently, the theoretical properties such as generalization error or free energy in the mean field approximation has been studied. The theoretical results give us the comparison with the regular statistical model and the foundation of a model selection. In this paper, we treat the complete bibpartite Boltzmann machines and derive the upper bound of asymptotic free energy of the mean field approximation., The Institute of Electronics, Information and Communication Engineers, 17 Mar. 2006, IEICE technical report, 105, 659, 125-130, Japanese, 0913-5685, 110004680405, AN10091178

Development of Ultrasound Robot for Automatic Acquisition of Ultrasound Images
Ryosuke Saito; Norihiro Koizumi; Yu Nishiyama; Tsubasa Imaizumi; Kenta Kusahara; Shiho Yagasaki; Masahiro Ogawa; Naoki Matsumoto
Oral presentation, English, 16th Annual Asian Conference on Computer Aided Surgery (ACCAS2020), International conference
27 Nov. 2020

Study on automatic aquisition of diagnostic images by ultrasound diagnostic robot
Yusuke Watanabe; Norihiro Koizumi; Yudai Sasaki; Kento Kobayashi; Takahiro Kobayashi; Zhou Jiayi; Yu Nishiyama; Hiroyuki Tsukihara; Kazushi Numata; Hideyuki Iijima; Toshiyuki Iwai; Hidetoshi Nagaoka
Oral presentation, English, 16th Annual Asian Conference on Computer Aided Surgery (ACCAS2020), International conference
27 Nov. 2020

Development of Ultrasound Robot for Automatic Acquisition of Ultrasound Images
Jiayi Zhou; Kento Kobayashi; Yusuke Watanabe; Takumi Fujibayashi; Momoko Matsuyama; Miyu Yamada; Hiroyuki Tsukihara; Yu Nishiyama; Norihiro Koizumi
Oral presentation, English, 16th Annual Asian Conference on Computer Aided Surgery (ACCAS2020)
27 Nov. 2020

Classification of hepatic hemangiomas and blood vessels from ultrasonography by deep learning
K. Kusahara; N. Koizumi; Y. Nishiyama; T. Imaizumi; R. Saito; S. Yagasaki; N. Matsumoto; M. Ogawa
Oral presentation, English, 34th International Congress and Exhibition on computer assisted radiology and surgery (CARS 2020), International conference
23 Jun. 2020

Contact state adjustment method to enhance organ motion compensation performance for a bed-type ultrasound diagnostic and therapeutic robot
K. Kobayashi; N. Koizumi; Y. Sasaki; T. Kobayashi; Y. Watanabe; J. Zhou; A. Otsuka; Y. Nishiyama; H.Tsukihara; N. Matsumoto; H. Miyazaki; K. Numata; H. Nagaoka; T. Iwai; H. Iijima
Oral presentation, English, 34th International Congress and Exhibition on computer assisted radiology and surgery (CARS 2020), International conference
23 Jun. 2020

Usefulness of computer-aided diagnosis system in evaluating severity of benign prostatic hyperplasia, using a super-ellipse model to characterize changes in prostate contours
S. Yagasaki; N. Koizumi; Y. Nishiyama; R. Kondo; T. Imaizumi; N. Matsumoto; M. Ogawa; K. Numata
Oral presentation, English, 34th International Congress and Exhibition on computer assisted radiology and surgery (CARS 2020), International conference
23 Jun. 2020

非侵襲超音波診断治療統合システムにおけるCNNを用いた患部追従手法に関する検証
五十嵐立樹; 冨田恭平; 小泉憲裕; 西山悠
Oral presentation, Japanese, 第25回ロボティクスシンポジア
15 Mar. 2020

医療診断・治療技能のデジタル化
小泉 憲裕; 西山 悠; 江浦 史生; 大塚 研秀; 佐々木 雄大; 重成 佑香; 五十嵐 立樹; 小林 賢人; 月原 弘之; 小路 直; 福田 浩之; 沼田 和司
Oral presentation, Japanese, 18回日本超音波治療研究会（JSTU2019), Domestic conference
07 Dec. 2019

Automatic Diagnosis by Compact Portable Ultrasound Robot : State Estimation of Internal Organs with Steady-State Kalman Filter
Yudai Sasaki; Fumio Eura; Kento Kobayashi; Ryosuke Kondo; Kyohei Tomita; Yu Nishiyama; Hiroyuki Tsukihara; Naoki Matsumoto; Norihiro Koizumi
Oral presentation, English, Proc. of 2019 IEEE Healthcare Innovations and Point of Care Technologies Conference (HI-POCT), International conference
20 Nov. 2019

ResNetを用いた超音波画像における肝血管腫と血管の抽出手法
草原 健太; 小泉 憲裕; 近藤 亮祐; 今泉 飛翔; 西山 悠; 松本 直樹; 小川 眞広
Oral presentation, Japanese, 日本超音波医学会第31回関東甲信越地方会学術集会, Domestic conference
20 Oct. 2019

画像診断支援システムをもちいた前立腺肥大症の 重症度予測の可能性
重成佑香; 小泉憲裕; 五十嵐立樹; 西山 悠; 小路 直
Oral presentation, Japanese, 第7回泌尿器画像診断・治療技術研究会（JSURT2019）, Domestic conference
14 Sep. 2019

Matching axial images of magnetic resonance imaging and transrectal ultrasound based on deep learning
R. Igarashi; N. Koizumi; Y. Nishiyama; K. Tomita; Y. Shigenari; S. Shoji
Oral presentation, English, Computer Assisted Radiology and Surgery (CARS) congress, International conference
18 Jun. 2019

Segmentation of liver parenchyma in ultrasound images for automatic diagnosis of liver cirrhosis
T. Imaizumi; N. Koizumi; Y. Nishiyama; N. Matsumoto; M. Ogawa
Oral presentation, English, Computer Assisted Radiology and Surgery (CARS) congress, International conference
18 Jun. 2019

アウトプットの重み付き和に基づくRNN言語モデル
川田航希; 西山悠; 川野秀一
Poster presentation, Japanese, 第20回情報論的学習理論ワークショップ(IBIS2017), Domestic conference
10 Nov. 2017

超音波治療モニタリングを支援する医デジ化コア基盤技術
小泉憲裕; 栢菅 篤; 細井泉澄; 冨田恭平; 近藤亮祐; 西山 悠; 月原弘之; 福田浩之; 葭中 潔; 東 隆; 宮嵜英世; 小路 直; 沼田和司; 松本洋一郎; 光石 衛
Oral presentation, Japanese, 第16回日本超音波治療研究会, Domestic conference
28 Oct. 2017

肝線維化の定量化を目的とするテクスチャ解析を用いたアーチファクトの分類手法
源田達也; 小泉憲裕; 大塚研秀; 近藤亮祐; 冨田恭平; 西山 悠; 坂無英徳; 熊川まり子; 松本直樹; 小川眞広
Oral presentation, Japanese, 日本超音波医学会第 90 回学術集会, Domestic conference
26 May 2017

ロバストかつ高精度な超音波ガイドRFA 治療支援システムの開発
近藤亮祐; 小泉憲裕; 冨田恭平; 西山 悠; 月原弘之; 福田浩之; 沼田和司; 光石 衛; 松本洋一郎
Oral presentation, Japanese, 日本超音波医学会第 90 回学術集会, Domestic conference
26 May 2017

非侵襲超音波診断治療統合システムのためのロバストかつ高精度な患部追従手法
冨田恭平; 小泉憲裕; 栢菅篤; 西山 悠; 月原弘之; 宮嵜英世; 葭仲 潔; 光石 衛
Poster presentation, Japanese, ロボティクス・メカトロニクス講演会2017, Domestic conference
10 May 2017

高次局所自己相関特徴を用いた動的テンプレートマッチングによる超音波ガイド腫瘍追従手法
近藤亮祐; 小泉憲裕; 冨田恭平; 西山悠; 福田浩之; 月原弘之; 沼田和司; 松本洋一郎; 光石 衛
Poster presentation, Japanese, ロボティクス・メカトロニクス講演会2017, Domestic conference
10 May 2017

医療用超音波のための音響シャドウを除去した臓器合成画像モデル
細井 泉澄; 小泉憲裕; 西山悠; 月原弘之; 宮嵜英世; 葭仲 潔; 光石衛
Poster presentation, Japanese, ロボティクス・メカトロニクス講演会2017, Domestic conference
10 May 2017

超音波画像における音響シャドウを除去した臓器合成モデルの構築法
細井泉澄; 小泉憲裕; 栢菅 篤; 冨田恭平; 西山 悠; 月原弘之; 福田浩之; 葭中 潔; 斎藤 季; 宮崎英世; 杉田直彦; 沼田和司; 本間之夫; 松本洋一郎; 光石 衛
Oral presentation, Japanese, 2017年度精密工学会春季大会学術講演会, Domestic conference
15 Mar. 2017

カーネルスペクトラルHMMを用いた風力予測
都築俊介; 西山悠
Poster presentation, Japanese, 新学術領域研究「スパースモデリングの深化と高次元データ駆動科学の創成」2016年度公開シンポジウム, Domestic conference
19 Dec. 2016

Robust servoing method for renal stones/tumors for the noninvasive ultrasound theragnostic system
Atsushi Kayasuga; Norihiro Koizumi; Kyohei Tomita; Yu Nishiyama; Hiroyuki Tsukihara; Hiroyuki Fukuda; Kiyoshi Yoshinaka; Takashi Azuma; Hideyo Miyazaki; Naohiko Sugita; Kazushi Numata; Yukio Honma; Yoichiro Matsumoto; Mamoru Mitsuishi
Oral presentation, English, The Journal of the Acoustical Society of America, International conference
01 Dec. 2016

Liver tracking system utilizing template matching and energy function in high intensity focused ultrasound/radio frequency ablation therapy
Kyohei Tomita; Norihiro Koizumi; Ryosuke Kondo; Atsushi Kayasuga; Yu Nishiyama; Hiroyuki Tsukihara; Hiroyuki Fukuda; Kazushi Numata; Yoichiro Matsumoto; Mamoru Mitsuishi
Oral presentation, English, The Journal of the Acoustical Society of America (JASA), International conference
01 Dec. 2016

An ultrasound guided monitoring system for high intensity focused ultrasound and radio frequency ablation therapies
Ryosuke Kondo; Norihiro Koizumi; Kyohei Tomita; Atsushi Kayasuga; Yu Nishiyama; Hiroyuki Tsukihara; Hiroyuki Fukuda; Kazushi Numata; Yoichiro Matsumoto; Mamoru Mitsuishi
Oral presentation, English, The Journal of the Acoustical Society of America (JASA), International conference
01 Dec. 2016

HIFU治療の高速・高精度化を実現する運動・変形する臓器の抽出・追従技術
栢菅篤; 小泉憲裕; 冨田恭平; 西山悠; 月原弘之; 福田浩之; 葭仲 潔; 東隆; 宮嵜英世; 杉田直彦; 沼田和司; 本間之夫; 松本洋一郎; 光石衛
Oral presentation, Japanese, 第15回日本超音波治療研究会, Domestic conference
12 Nov. 2016

肝線維化の定量化を目的とする血管等アーチファクトの除去手法
源田達也; 小泉憲裕; 近藤亮祐; 江浦史生; 西山悠; 熊川まり子; 松本直樹; 小川眞広
Oral presentation, Japanese, 日本超音波医学会 第28回関東甲信越地方会学術集会, Domestic conference
22 Oct. 2016

超音波B-Flow画像を用いた肝硬変診断のための血管分岐モデルおよび解析
近藤亮祐; 小泉憲裕; 源田達也; 江浦史生; 西山悠; 熊川まり子; 松本直樹; 小川眞広
Oral presentation, Japanese, 日本超音波医学会 第28回関東甲信越地方会学術集会, Domestic conference
22 Oct. 2016

McKibben型空気圧ゴム人工筋の製品種別に関する識別器の構成
石川貴大; 岡部篤; 西山悠; 小木曽公尚
Oral presentation, Japanese, 第3回計測自動制御学会 制御部門マルチシンポジウム(MSCS2016)
07 Mar. 2016

Nonparametric Kernel Bayes Smoothing on State Space Models
西山悠
Poster presentation, Japanese, 新学術領域研究「スパースモデリングの深化と高次元データ駆動科学の創成」2015年度公開シンポジウム, Domestic conference
07 Mar. 2016

kNNを用いたカーネルベイズの計算量削減法の検討
苗村智行; 都築俊介; 西山悠
Poster presentation, Japanese, 第18回情報論的学習理論ワークショップ(IBIS2015), Domestic conference
26 Nov. 2015

カーネルベイズスムージングとカーネル平均Toolboxの作成
西山悠
Oral presentation, Japanese, 第25回日本神経回路学会全国大会(JNNS 2015)
04 Sep. 2015

Nonparametric Smoothing on State Space Models with Kernel Mean Embeddings
Yu Nishiyama; Amir Hossein Afsharinejad; Shunsuke Naruse; Byron Boots; Le Song
Poster presentation, English, 1st Symposium on Intelligent Systems in Science and Industry (SISSI)
12 Jul. 2015

Learning Bayesian networks: Recursive autonomy identification algorithm incorporating a strict learning
K. Natori; M. Uto; Y. Nishiyama; S Kawano; M Ueno
Poster presentation, Japanese, 数学協働プログラム「確率的グラフィカルモデル」, Domestic conference
19 Mar. 2015

カーネル平均によるカーネルベイズ推論と確率モデルの融合
西山悠
Public discourse, Japanese, 山梨大学医学部キャンパス解剖学講座細胞生物学教室, Invited
12 Feb. 2015

Model-based Kernel Sum Rule with Applications to State Space Models
Yu Nishiyama; Motonobu Kanagawa; Arthur Gretton; Kenji Fukumizu
Oral presentation, English, The Neural Information Processing Systems (NIPS) Workshop, Invited, International conference
12 Dec. 2014

カーネル法と確率分布の無限分解可能性
西山 悠
Invited oral presentation, Japanese, 日本応用数理学会 2014年度年会, Domestic conference
05 Sep. 2014

カーネル平均を使ったカーネルベイズ推論と無限分解可能過程の交錯に向けて
西山 悠; 統計数理研究所
Public discourse, Japanese, 九大セミナー, 福岡 日本
17 Mar. 2014

特性的カーネルと畳み込み無限分解可能カーネル
西山 悠; 統計数理研究所; 福水 健次; 統計数理研究所
Oral presentation, Japanese, 第8回日本統計学会春季集会, 京都 日本
08 Mar. 2014

最近のカーネル法として正定値カーネルを使ったベイズ推論と確率モデルとの融合
西山 悠; 統計数理研究所
Public discourse, Japanese, グラフマイニング＆WEB＆AIセミナー, 東京 日本
20 Jan. 2014

最近のカーネル法として正定値カーネルを使ったベイズ推論の話題
西山 悠; 統計数理研究所
Public discourse, Japanese, インシリコ・メガバンク研究会, 仙台 日本
10 Dec. 2013

無限分解可能分布におけるカーネル平均の検討
西山 悠; 福水 健次
Public symposium, Japanese, 第16回情報論的学習理論ワークショップ(IBIS2013), 東京 日本, Domestic conference
13 Nov. 2013

パラメトリックカーネル平 均を用いた状態空間フィルタリングアルゴリズム
西山 悠; 金川 元信; Arthur Gretton; 福水 健次
Public symposium, Japanese, 第15 回情報論的学習理論ワークショップ(IBIS2012), 東京 日本
08 Nov. 2012

カーネル法によるパーティ クルフィルタ
金川 元信; 西山 悠; Arthur Gretton; 福水 健次
Public symposium, Japanese, 第15 回情報論的学習理論ワークショップ(IBIS2012), 東京 日本
08 Nov. 2012

Kernel Mean Embeddings of POMDPs
Nishiyama, Yu; 統計数理研究所; Boularias, Abdeslam; マックスプランク研究所; Gretton, Arthur; University College London; Fukumizu; Kenji; 統計数理研究
Poster presentation, English, Machine Learning Summer Schools 2012, 京都 日本
28 Aug. 2012

部分観測マルコフ決定過程ベルマン方程式のカーネル化
西山 悠; BOULARIAS Abdeslam; GRETTON Arthur; 福水 健次
Oral presentation, Japanese, 電子情報通信学会技術研究報告. IBISML, 情報論的学習理論と機械学習
05 Mar. 2012

Probabilistic Inference by minimizing the TRW Free Energy using CCCP
NISHIYAMA Yu; YE Xingyao; YUILLE lan L
Poster presentation, Japanese, 電子情報通信学会技術研究報告. IBISML, 情報論的学習理論と機械学習, The Institute of Electronics, Information and Communication Engineers, We propose a family of convergent double-loop algorithms which minimize the TRW free energy. These algorithms are based on the concave convex procedure (CCCP) so we call them TRW-CCCP. Our formulation includes many free parameters which specify an infinite number of decompositions of the TRW free energy into convex and concave parts. TRW-CCCP is guaranteed to converge to the global minima for any settings of these free parameters, including adaptive settings if they satisfy conditions denned in this paper. We show that the values of these free parameters control the speed of convergence of the the inner and outer loops in TRW-CCCP. We performed experiments on a two-dimensional Ising model observing that TRW-CCCP converges to the global minimum of the TRW free energy and that the convergence rate depends on the parameter settings. We compare with the original message passing algorithm (TRW-BP) by varying the difficulty of the problem (by adjusting the energy function) and the number of iterations in the inner loop of TRW-CCCP. We show that on difficult problems TRW-CCCP converges faster than TRW-BP (in terms of total number of iterations) if few inner loop iterations are used.
28 Oct. 2010

A CDMA Multiuser Demodulation Algorithm Based on NCCCP
NISHIYAMA Yu; TONOSAKI Yukinori; WATANABE Sumio
Oral presentation, Japanese, IEICE technical report. Neurocomputing, The Institute of Electronics, Information and Communication Engineers, For CDMA demodulation problem in mobile communication technology, analyses of performance and developments of algorithms on the basis of statistical mechanical informatics have been done. While CDMA multiuser demodulation algorithm based on belief propagation(BP) shows its efficacy especially in large systems with respect to users and chips, that based on CCCP is known to provide better performance in small systems, which often appear in practical situation. However, CCCP has a drawback that it requires huge computational cost. Recently, we proposed a new CCCP(NCCCP) algorithm, which is an extension of CCCP algorithm for the Bethe free energy and show it can reduce expensive computational cost. In this report, we develop the CDMA multiuser demodulation algorithm based on NCCCP and compare the performance with that of conventional CCCP.
13 Dec. 2008

An Extension of CCCP Algorithm for Bethe Free Energy
NISHIYAMA Yu; WATANABE Sumio
Oral presentation, Japanese, IEICE technical report. Neurocomputing, The Institute of Electronics, Information and Communication Engineers, Belief Propagation (BP) is an efficient algorithm for computing marginal probabilities of a high-dimensional probability distribution. The marginals computed by BP are equivalent to the extrema of Bethe free energy. Concave convex procedure (CCCP) has been studied for the optimization of Bethe free energy. In this paper, we extend the CCCP algorithm for Bethe free energy and present a new CCCP (NCCCP) algorithm. We practically apply NCCCP algorithm to multi-dimensional Gaussian distributions. As a result, NCCCP algorithm enables inner loop to converge even if all Lagrange multipliers in the loop are simultaneously updated. Moreover, we find that there exists an optimal point of the parameters introduced to NCCCP that can reduce the expensive computational cost.
05 Mar. 2008

25pPSB-4 An Extension of CCCP Algorithm for Kikuchi Free Energy
Nishiyama Yu; Watanabe Sumio
Poster presentation, Japanese, Meeting abstracts of the Physical Society of Japan, The Physical Society of Japan (JPS)
29 Feb. 2008

Analytical and Numerical Solutions of Bethe Approximation in Normal Distributions
NISHIYAMA Yu; WATANABE Sumio
Oral presentation, Japanese, IEICE technical report. Neurocomputing, The Institute of Electronics, Information and Communication Engineers, For the calculation of marginal distributions, which require the huge computational cost, the algorithm called belief propagation(BP) has shown the effectiveness. Loopy belief propagation(LBP), which is the BP applied to the distributions that have loops, is not guaranteed to converge in general and, if it converges, it computes approximate marginal probabilities. The LBP fixed-points after the convergence are also given by the extrema of the evaluation function called Bethe free energy. In this paper, we clarify the extrema of Bethe free energy in some particular cases of normal distributions, and after we give the convergence condition of LBP and the accuracy of LBP, we do the numerical experiments and compare the results with analytical solutions.
08 Jan. 2008

Theoretical Analysis of Accuracy of Belief Propagation in Gaussian Models
NISHIYAMA Yu; WATANABE Sumio
Oral presentation, Japanese, IEICE technical report. Neurocomputing, The Institute of Electronics, Information and Communication Engineers, Belief propagation (BP) is and algorithm which can compute marginal probability distributions with a tractable computational cost. Loopy belief propagation (LBP) applied to the graphs containing loops is known to provide marginal distributions approximately if LBP converges. In this paper, we apply LBP to a multi-dimensional Gaussian distribution that has loops and analytically show how accurate LBP is in some cases. Specifically, we analytically show messages, approximate marginal densities, and the KL distance at fixed points of LBP when the graph corresponding to a Gaussian distribution has at most a single loop. Basides, for the graphs which have arbitrary structures, we derive the expansions of approximate marginal densities when covariances are small.
14 May 2007

Asymptotic Behavior of Free Energy of General Boltzmann Machines in Mean Field Approximation
NISHIYAMA Yu; WATANABE Sumio
Oral presentation, Japanese, IEICE technical report. Neurocomputing, The Institute of Electronics, Information and Communication Engineers, In the Bayesian learning, which generally requires huge computational costs, the algorithms based on the mean field approximation have shown us the effectiveness in the practical information systems. Recently, the generalization error or free energy in the mean field approximation has been theoretically studied. The theoretical results enable us to know the accuracy of the approximation and contribute to the foundation of a model selection in statistical singular machines. In this paper, we show that the upper bounds of the asymptotic free energies are theoretically obtained by counting the number of non-0 eigenvalues of Fisher information matrices and derive the upper bound in the learning model of general Boltzmann machines.
07 Jul. 2006

Stochastic Complexity of Complete Bibpartite Graph-type Boltzmann Machines in Mean Field Approximation
NISHIYAMA Yu; WATANABE Sumio
Oral presentation, Japanese, IEICE technical report. Neurocomputing, The Institute of Electronics, Information and Communication Engineers, In the learning of singular learning machines, the superiority of Bayesian learning is shown. However, it requires huge computational costs to realize the Bayesian a posteriori distribution. To overcome this problem, the mean field approximation, which is originally known in statistical physics, is used in the practical information systems. Recently, the theoretical properties such as generalization error or free energy in the mean field approximation has been studied. The theoretical results give us the comparison with the regular statistical model and the foundation of a model selection. In this paper, we treat the complete bibpartite Boltzmann machines and derive the upper bound of asymptotic free energy of the mean field approximation.
10 Mar. 2006

Advanced Statistical Machine Learning
Oct. 2023 - Present
The University of Electro-Communications

Applied Mathematics Ⅰ
Oct. 2022 - Present
The University of Electro-Communications

情報数理工学実験第二B
Oct. 2019 - Present
The University of Electro-Communications

Exercises in Information and Communication Engineering II
Apr. 2018 - Sep. 2023
The University of Electro-Communications

Exercises in Information and Communication Engineering I
Apr. 2016 - Sep. 2023
The University of Electro-Communications

Graduate Technical English
Apr. 2021 - Jul. 2023
The University of Electro-Communications

K課程輪講
Apr. 2020 - Mar. 2021
The University of Electro-Communications

Exercise in Informatics I
Oct. 2019 - Mar. 2021
The University of Electro-Communications

Information Engineering Laboratory
Apr. 2016 - Mar. 2020
The University of Electro-Communications

Innovative Comprehensive Communications Design 1
Apr. 2018 - Sep. 2018
The University of Electro-Communications

社会知能情報学基礎1
Apr. 2015 - Sep. 2016
The University of Electro-Communications

合同輪講
Oct. 2014 - Mar. 2016
The University of Electro-Communications

Construction of robotic ultrasound diagnostic and treatment platform system to accelerate and promote Medical DigITalization (Me-DigIT)
小泉 憲裕; 小木曽 公尚; 月原 弘之; 西山 悠; 宮嵜 英世
Japan Society for the Promotion of Science, Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B), The University of Electro-Communications, Grant-in-Aid for Scientific Research (B), 本研究課題の目的は、人工知能技術・ロボット技術・超音波技術を用いてロボティック超音波診断・治療基盤システムの構築法を確立することであり、下記の5つのコア基盤技術に関する研究を遂行し、下記に示すようにきわめて順調に優れた成果を積み重ねてきている。(コア技術I) 機能に応じた機構設計技術、 (コア技術II) 医療診断・治療技能における機能の抽出・構造化技術、 (コア技術III) 患者に対するロボットの安全・安心動作技術、(コア技術IV) 診断・治療タスクに応じたシステム動作技術、(コア技術V) リアルタイム医用画像処理技術。 とりわけ (コア技術V) リアルタイム医用画像処理技術については超音波画像中に表示される臓器が呼吸や拍動に伴って変位・変形・回転する場合や音響陰影によって画像が一部欠損した場合にも、特定臓器の輪郭や患部を画像合成および抽出・追従・モニタリングできる、深層学習を援用した画像処理技術を新規に開発するなど、成果を順調に積み重ねてきている（第22回日本超音波医学会奨励賞受賞）。また、(コア技術I)に関してロボティック超音波医療診断・治療支援システム（アラベスク）の有効性を評価するために診断対象であるファントムの体位・姿勢を制御する、患者体位・姿勢制御ロボット（ピルエット）を新規開発した。これにより、ロボティック超音波診断システムが指定された位置・姿勢・接触力で超音波プローブをアプローチして、任意の臓器内の患部に対して超音波診断画像を獲得することが可能になった。 ほかにもAI・ロボティック支援医療診断・治療システム分野の一流国際誌（IJCARS）への論文掲載、トップカンファレンスでの発表（CARS2021）、日経新聞等への掲載など，医療診断・治療のための生体患部抽出・追従・モニタリング技術のパイオニアかつ中核的な存在として国内外からきわめて高い注目を集めてきている。, 20H02113
01 Apr. 2020 - 31 Mar. 2024

カーネルベイズ推論に基づく時系列アルゴリズムの開発と展開
西山 悠
日本学術振興会, 科学研究費助成事業 基盤研究(C), 電気通信大学, 基盤研究(C), 【背景】状態空間モデルによる時系列モデリングでは，フィルタリングアルゴリズムと平滑化アルゴリズムの開発が重要である．離散時間時不変状態空間モデルの設定として，観測過程の条件付き確率をカーネル平均埋め込み手法でノンパラメトリックに推定し，状態遷移過程の条件付き確率をパラメトリックモデルに推定する状況を考える．このとき，model-based kernel Bayes’ filter (Mb-KBF)が提案されている． 【問題点とそれに対する提案法】 1.Mb-KBFに対応する平滑化アルゴリズムmodel-based kernel Bayes’ smoother (Mb-KBS)は未開発であった．そこでMb-KBSアルゴリズムを開発し，Stochastic Volatility モデルの場合に有効性を検証した．関連手法 (Mb-KBF, nonparametric kernel Bayes smoothing)と比較して隠れ状態の推定精度 (RMSE)が高くなる結果を得た．本アルゴリズムは4つの超パラメータのチューニングを必要とするが，超パラメータを変化させたときのアルゴリズムの振る舞いの詳細な検証を行い，また直観的に分かりやすい可視化を行った． 2.状態遷移過程時に大きな外れ値ノイズが発生する場合，加法的ガウスノイズモデルを用いた学習より，加法的コーシーノイズモデルを用いた学習が有効と考えられる．しかしそのアルゴリズムは未開発であった．そこで状態遷移過程時に大きな外れ値ノイズが発生する状況に対応するため，加法的コーシーノイズモデルとコーシーカーネルの共役性を組み合わせたMb-KBFを開発した．またコーシーカーネルのときに点推定アルゴリズムを開発した．数値実験の結果，提案手法は加法的ガウスノイズモデルを用いた学習より，隠れ状態の推定精度 (RMSE)が高くなる結果を得た．, 20K11933
01 Apr. 2020 - 31 Mar. 2023

ビッグデータ解析による地域医療の実態解明
藤林和俊
公益財団法人大樹生命厚生財団, 第52回医学研究助成
01 Apr. 2019 - 31 Mar. 2021

Model-based control and management technology for pneumatic artificial muscle actuators
Kogiso Kiminao
Japan Society for the Promotion of Science, Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (C), The University of Electro-Communications, Grant-in-Aid for Scientific Research (C), This study has obtained four results related to model-based control, management, and securing of pneumatic artificial muscles (PAM) that realize flexible actuators. Specifically, we have identified a precise mathematical model of the PAM. We have developed model-based control and sensorless control to track the desired flexibility (stiffness), a method for detecting the failure of the PAM by focusing on changes in model parameters of the PAM model, and a secure control method by applying encrypted control., 18K04012
01 Apr. 2018 - 31 Mar. 2021

Ultrahigh accuracy of ultrasound theragnostic systems realized by technologizing and digitizing medical professional skills
KOIZUMI NORIHIRO
Japan Society for the Promotion of Science, Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B), The University of Electro-Communications, Grant-in-Aid for Scientific Research (B), The purpose of this research is to establish a method for constructing an integrated ultrasonic diagnosis / treatment system that operates robustly and with high accuracy, and excellent research results regarding the core basic technology related to robot mechanism / control / image processing algorithm technology for this purpose. Have been piled up. In particular, medical image processing technology has made epoch-making breakthroughs in the field of image processing by machine learning such as deep learning in recent years, and this project also started research on medical robot vision technology that incorporates deep learning. The results have been steadily accumulating (received the 25th Robotics Symposia Student Encouragement Award (2020), the 21st and 22nd Japan Society of Ultrasonics Medicine Encouragement Award (2020, 2021), etc.)., 17H03200
01 Apr. 2017 - 31 Mar. 2021

情報通信技術(機械学習/人工知能)による糖尿病臨床支援システム開発
Novartis Research Grants
01 Apr. 2018 - 31 Mar. 2019

Deepening and applications of sparse modeling by approaches of semiparametric Bayesian inference
Fukumizu Kenji; NISHIYAMA Yu; LIU Song
Japan Society for the Promotion of Science, Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research on Innovative Areas (Research in a proposed research area), The Institute of Statistical Mathematics, Grant-in-Aid for Scientific Research on Innovative Areas (Research in a proposed research area), Filtering problems aims at estimating the current unobserved state variable from the unknown dynamics of the unobserved state variables and indirect observations. We consider filtering under the assumption that the observation model is uncertain and not able to be modeled easily, and proposed effective algorithms in such difficult situations. We confirmed the advantage of the proposed algorithms over existing relevant methods. We also studied fast methods for complex sparse modeling, and proposed an optimization methods that achieves the best convergence rate theoretically., 25120012
28 Jun. 2013 - 31 Mar. 2018

Kernel Bayes Inference and Infinitely Divisible Distributions
Nishiyama Yu
Japan Society for the Promotion of Science, Grants-in-Aid for Scientific Research Grant-in-Aid for Young Scientists (B), The University of Electro-Communications, Grant-in-Aid for Young Scientists (B), Principal investigator, Kernel Bayes Inference (KBI), which is a Bayesian inference based on kernel methods, has been studied. KBI infers kernel means, which are features of probability distributions in reproducing kernel Hilbert space. In KBI, characteristic kernels play an important role in specifying probability distributions by kernel means. We studied a connection between characteristic kernels and infinitely divisible distributions. We showed that continuous bounded and symmetric density functions of infinitely divisible distributions can be used for characteristic kernels. Within the infinitely divisible distributions, we proposed a convolution trick, which is a generalization of the kernel trick. The convolution trick can be used for developing various kernel algorithms that combine infinitely divisible distributions., 26870821
01 Apr. 2014 - 31 Mar. 2017

確率伝播法の理論解析とその設計法
西山悠
日本学術振興会, 特別研究員奨励費(課題番号: 07J05165), 東京工業大学, 特別研究員奨励費, Principal investigator, 確率伝播法に代表されるメッセージパシングアルゴリズムは、機械故障診断、人工知能、人間行動モデリング、コンピュータビジョン、システムバイオロジー、脳情報処理を始めとして幅広い応用を持つ。確率伝播法は、ループを持つグラフ上で定義された確率分布に適用した場合、収束保証のない近似アルゴリズムとなることが知られるが、広く使われる多次元正規分布に適用した場合、近似精度と収束条件を解析的に明らかにした研究について、今年度、論文発表を行った。また収束が保証されるメッセージパシングアルゴリズムに、CCCP法があるが、これを拡張したnew CCCP法を昨年度から提案している。今年度は、CCCPの提案者であるアメリカUCLA大学のAlan L. Yuille教授の研究室を訪問し、CCCP法についての共同研究を行った。マルコフ確率場における近似推論の研究で、近年TRW自由エネルギーが注目を集めている。Yuille教授との共同研究の結果、TRW自由エネルギーをCCCP法によって最小化するアルゴリズム(TRW-CCCP)を開発した。確率分布において、最大確率を与える状態を求める問題はMAP問題として知られるが,整数計画法を線形計画緩和し、メッセージパシングにより効率的に近似計算するアルゴリズム研究が盛んに行われている。提案したTRW-CCCPは、MAP計算も行うことができるアルゴリズムである。またTRW自由エネルギーの双対エネルギーを求め、それを最適化することで、より効率的なアルゴリズムの開発を行っている。与えられた行列から二重確率行列を求めるアルゴリズムにSinkhornアルゴリズムが知られるが、これはCCCP法として解釈可能であることが知られる。この拡張アルゴリズムをnew CCCP法に基づいて与えた。これらの一連の研究について、これから、国際会議や論文誌での発表を行う予定である。, 07J05165
Apr. 2007 - 2009

超音波治療診断システム、超音波治療診断方法、プログラムおよび患部追従評価方法
Patent right, 小泉 憲裕, 大塚 研秀, 西山 悠, 月原 弘之, 宮嵜 英世, 沼田 和司, 特願2021-069893, Date applied: 16 Apr. 2021, 国立大学法人電気通信大学, 国立大学法人 東京大学, 国立研究開発法人国立国際医療研究センター, 特開2022-164416, Date announced: 27 Oct. 2022

生体内運動追跡装置、生体内運動追跡装置の作動方法およびプログラム
Patent right, 小泉 憲裕, 西山 悠, 近藤 亮祐, 冨田 恭平, 江浦 史生, 沼田 和司, 特願2017-158071, Date applied: 18 Aug. 2017, 国立大学法人電気通信大学, 公立大学法人横浜市立大学, 株式会社大林製作所, 特開2019-033960, Date announced: 07 Mar. 2019, 特許第7037136号, Date registered: 08 Mar. 2022

生体内運動追跡装置
Patent right, 小泉 憲裕, 栢菅 篤, 冨田 恭平, 細井 泉澄, 西山 悠, 月原 弘之, 宮嵜 英世, 福田 浩之, 沼田 和司, 葭仲 潔, 東 隆, 杉田 直彦, 本間 之夫, 松本 洋一郎, 光石 衛, 特願2017-040348, Date applied: 03 Mar. 2017, 国立大学法人 東京大学, 国立大学法人電気通信大学, 公立大学法人横浜市立大学, 特開2018-143416, Date announced: 20 Sep. 2018, 特許第6829437号, Date registered: 26 Jan. 2021

生体内運動追跡装置、生体内運動追跡方法およびプログラム
Patent right, 小泉 憲裕, 西山 悠, 近藤 亮祐, 冨田 恭平, 江浦 史生, 沼田 和司, 特願2017-158071, Date applied: 18 Aug. 2017, 国立大学法人電気通信大学, 公立大学法人横浜市立大学, 特開2019-033960, Date announced: 07 Mar. 2019

生体内運動追跡装置
Patent right, 小泉 憲裕, 栢菅 篤, 冨田 恭平, 細井 泉澄, 西山 悠, 月原 弘之, 宮嵜 英世, 福田 浩之, 沼田 和司, 葭仲 潔, 東 隆, 杉田 直彦, 本間 之夫, 松本 洋一郎, 光石 衛, 特願2017-040348, Date applied: 03 Mar. 2017, 国立大学法人 東京大学, 国立大学法人電気通信大学, 公立大学法人横浜市立大学, 国立研究開発法人産業技術総合研究所, 特開2018-143416, Date announced: 20 Sep. 2018

文科省委託事業「数学・数理科学と諸科学・産業との協働によるイノベーション創出のための研究促進プログラム」数理・材料科学ワーキンググループ, 協力研究者
2013 - 2013