東北大学研究者紹介

経歴 3

2025年4月～継続中

東北大学　大学院情報科学研究科情報セキュリティ論講座　助教
2024年3月～継続中

東北大学　データ駆動科学・AI教育研究センターデータ基盤・セキュリティ教育研究部門　助教
2023年11月～ 2024年2月

東北大学医学部医用画像工学分野　学術研究員

学歴 3

東北大学　大学院医工学研究科

2018年10月～ 2023年9月
東北大学　大学院医工学研究科

2016年10月～ 2018年9月
北京航空航天大学　高等工程学院

2012年9月～ 2016年6月

委員歴 1

計測自動制御学会システム・情報部門学術講演会（SSI2025）　現地実行委員（会場担当）

2025年～ 2025年

所属学協会 1

IEEE

2023年12月～継続中

研究キーワード 3

説明可能AI
医用画像診断支援システム
人工知能

研究分野 2

情報通信 / 知覚情報処理 /
ライフサイエンス / 医用システム /

受賞 2

The Best Paper Prize

2023年12月　IEEE Sendai Section
SSI Excellent Paper Award

2018年11月　SICE System and Information Division

論文 11

Transformer-based Deep Learning Models with Shape Guidance for Predicting Breast Cancer in Mammography Images 査読有り

Kengo Takahashi, Yuwen Zeng, Zhang Zhang, Kei Ichiji, Takuma Usuzaki, Ryusei Inamori, Haoyang Liu, Noriyasu Homma

Journal of Imaging Informatics in Medicine　2025年12月19日
出版者・発行元： Springer Science and Business Media LLC
DOI： 10.1007/s10278-025-01773-3 　

eISSN：2948-2933

詳細を見る詳細を閉じる

Abstract Recent breast cancer research has investigated shape-based attention guidance in Vision Transformer (ViT) models, focusing on anatomical structures and the heterogeneity surrounding tumors. However, few studies have clarified the optimal transformer encoder layer stage for applying attention guidance. Our study aimed to evaluate the effectiveness of shape-guidance strategies by varying the combinations of encoder layers that guide attention to breast structures and by comparing the proposed models with conventional models. For the shape-guidance strategy, we applied breast masks to the attention mechanism to emphasize spatial dependencies and enhance the learning of positional relationships within breast anatomy. We then compared the representative models—Masked Transformer models that demonstrated the best performance across layer combinations—with the conventional ResNet50, ViT, and SwinT V2. In our study, a total of 2,436 publicly available mammography images from the Chinese Mammography Database via The Cancer Imaging Archive were analyzed. Three-fold cross-validation was employed, with a patient-wise split of 70% for training and 30% for validation. Model performance on differentiating breast cancer from non-cancer images was assessed by the area under the receiver-operating characteristic curve (AUROC). The results showed that applying masks at the Shallow and Deep stages gave the highest AUROC for Masked ViT. The Masked ViT achieved an AUROC of 0.885 [95% confidence interval: 0.849–0.918], a sensitivity of 0.876, and a specificity of 0.802, outperforming all other conventional models. These results indicate that incorporating mask guidance into particular Transformer encoders promotes representation learning, highlighting their potential as decision-support tools in breast cancer diagnosis.
Robust Ownership Verification of DNN Models Against JPEG Compression via Probability-Controlled Adversarial Attacks

Teruki Sano, Minoru Kuribayashi, Masao Sakai, Shuji Isobe, Eisuke Koizumi, Zhang Zhang

2025 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)　2140-2145　2025年10月22日
出版者・発行元： IEEE
DOI： 10.1109/apsipaasc65261.2025.11249003 　
Adaptive region-oriented masked vision retentive network for predicting macrovascular invasion in hepatocellular carcinoma 査読有り

Kengo Takahashi, Ryusei Inamori, Kei Ichiji, Zhang Zhang, Zeng Yuwen, Noriyasu Homma

Medical Imaging 2025: Computer-Aided Diagnosis　65-65　2025年4月4日
出版者・発行元： SPIE
DOI： 10.1117/12.3045625 　
Bilateral Information-Guided Diagnosis of Breast Masses in Mammography Using Vision Transformer 査読有り

Tianyu Zeng, Yuwen Zeng, Zhang Zhang, Xiaoyong Zhang, Kei Ichiji, Shuo-Yan Chou, Ivo Bukovsky, Jan Vrba, Noriyasu Homma

IEEE Journal of Biomedical and Health Informatics　1-13　2025年
出版者・発行元： Institute of Electrical and Electronics Engineers (IEEE)
DOI： 10.1109/jbhi.2025.3626200 　

ISSN：2168-2194

eISSN：2168-2208
Vision Transformer-Based Breast Mass Diagnosis in Mammography Using Bilateral Information 査読有り

Tianyu Zeng, Zhang Zhang, Yuwen Zeng, Xiaoyong Zhang, Kei Ichiji, Noriyasu Homma

2024 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)　147-152　2024年8月26日
出版者・発行元： IEEE
DOI： 10.1109/iicaiet62352.2024.10730097 　
Attention Optimization in AI-Aided Drowning Diagnosis Using Post-Mortem CT to Mitigate Overfitting with Limited Training Data 査読有り

Zhang Zhang, Xiaoyong Zhang, Taihei Mizuno, Kei Ichiji, Noriyasu Homma

2024 International Joint Conference on Neural Networks (IJCNN)　32　1-6　2024年6月30日
出版者・発行元： IEEE
DOI： 10.1109/ijcnn60899.2024.10650327 　
How intra-source imbalanced datasets impact the performance of deep learning for COVID-19 diagnosis using chest X-ray images 査読有り

Zhang Zhang, Xiaoyong Zhang, Kei Ichiji, Ivo Bukovský, Noriyasu Homma

Scientific Reports　13　(1)　2023年11月3日
出版者・発行元： Springer Science and Business Media LLC
DOI： 10.1038/s41598-023-45368-w 　

eISSN：2045-2322

詳細を見る詳細を閉じる

Abstract Over the past decade, the use of deep learning has been widely increasing in the medical image diagnosis field. Deep learning-based methods’ (DLMs) performance strongly relies on training data. Therefore, researchers often focus on collecting as much data as possible from different medical facilities or developing approaches to avoid the impact of inter-category imbalance (ICI), which means a difference in data quantity among categories. However, due to the ICI within each medical facility, medical data are often isolated and acquired in different settings among medical facilities, known as the issue of intra-source imbalance (ISI) characteristic. This imbalance also impacts the performance of DLMs but receives negligible attention. In this study, we study the impact of the ISI on DLMs by comparison of the version of a deep learning model that was trained separately by an intra-source imbalanced chest X-ray (CXR) dataset and an intra-source balanced CXR dataset for COVID-19 diagnosis. The finding is that using the intra-source imbalanced dataset causes a serious training bias, although the dataset has a good inter-category balance. In contrast, the deep learning model performed a reliable diagnosis when trained on the intra-source balanced dataset. Therefore, our study reports clear evidence that the intra-source balance is vital for training data to minimize the risk of poor performance of DLMs.
How Different Data Sources Impact Deep Learning Performance in COVID-19 Diagnosis using Chest X-ray Images 査読有り

Zhang Zhang, Xiaoyong Zhang, Kei Ichiji, Ivo Bukovský, Shuoyan Chou, Noriyasu Homma

2023 14th IIAI International Congress on Advanced Applied Informatics (IIAI-AAI)　4　508-513　2023年7月8日
出版者・発行元： IEEE
DOI： 10.1109/iiai-aai59060.2023.00103 　
Risk Analysis of Breast Cancer by Using Bilateral Mammographic Density Differences: A Case-Control Study 査読有り

Zhang Zhang, Xiaoyong Zhang, Jiaqi Chen, Yumi Takane, Satoru Yanagaki, Naoko Mori, Kei Ichiji, Katsuaki Kato, Mika Yanagaki, Akiko Ebata, Minoru Miyashita, Takanori Ishida, Noriyasu Homma

The Tohoku Journal of Experimental Medicine　261　(2)　139-150　2023年
出版者・発行元： Tohoku University Medical Press
DOI： 10.1620/tjem.2023.j066 　

ISSN：0040-8727

eISSN：1349-3329
Deep Learning-Based Diagnosis of Fatal Hypothermia Using Post-Mortem Computed Tomography 査読有り

Yuwen Zeng, Xiaoyong Zhang, Issei Yoshizumi, Zhang Zhang, Taihei Mizuno, Shota Sakamoto, Yusuke Kawasumi, Akihito Usui, Kei Ichiji, Ivo Bukovsky, Masato Funayama, Noriyasu Homma

The Tohoku Journal of Experimental Medicine　260　(3)　253-261　2023年
出版者・発行元： Tohoku University Medical Press
DOI： 10.1620/tjem.2023.j041 　

ISSN：0040-8727

eISSN：1349-3329
Adaptive Gaussian Mixture Model-Based Statistical Feature Extraction for Computer-Aided Diagnosis of Micro-Calcification Clusters in Mammograms 査読有り

Zhang Zhang, Xiaoyong Zhang, Kei Ichiji, Yumi Takane, Satoru Yanagaki, Yusuke Kawasumi, Tadashi Ishibashi, Noriyasu Homma

SICE Journal of Control, Measurement, and System Integration　13　(4)　183-190　2020年7月1日
出版者・発行元： Informa UK Limited
DOI： 10.9746/jcmsi.13.183 　

ISSN：1882-4889

eISSN：1884-9970

︎全件表示 ︎最初の5件までを表示

MISC 6

LLMを用いた放射線読影レポートの平易化における語彙頻度に基づく可読性評価の検討

曾昱雯, 三上真理子, 張彰, 張暁勇, 本間経康

2026年2月
逆変形ベクトル場表現を用いた 3 次元体内動態の粒子フィルタに基づく推定

川浪佑介, 市地慶, 張彰, 本間経康

SSI2025　2025年11月
High-Resolution Abnormal Mammogram Generation from Normal Samples Using Density-Aware Dual-Mask CycleGAN for Imbalanced Dataset Augmentation

Wuhong Jiang, Zhang Zhang, Yuwen Zeng, Kei Ichiji, Xiaoyong Zhang, Noriyasu Homma

SSI2025　2025年11月
Adversarial domain adaptation for breast cancer diagnosis in multi-center datasets

Xinyang He, Yuwen Zeng, Zhang Zhang, Kei Ichiji, Xiaoyong Zhang, Noriyasu Homma

SSI2025　2025年11月
深層学習は乳癌画像をどう読むか

本間経康, 本間経康, 張暁勇, 張暁勇, 高野寛己, 野呂恭平, 張彰, 陳家旗, 市地慶, 市地慶, 杉田典大, 酒井正夫, 吉澤誠, 川住祐介, 石橋忠司

日本乳癌画像研究会プログラム・抄録集　28th　2019年
Computer-Aided Diagnosis of Micro-Calcification Clusters in Mammograms Using an Adaptive Gaussian Mixture Model

Zhang, Z, Zhang, X, Ichiji, K., Osanai, M., Homma, N.

SSI2018　2018年11月

︎全件表示 ︎最初の5件までを表示

講演・口頭発表等 6

データの施設多様性が乳房腫瘤検出モデルの性能に与える影響の分析

菅野真梨子, 張彰, Yuwen Zeng, 市地慶, 張暁勇, 本間経康

第53回日本放射線技術学会秋季学術大会　2025年
局所定速度運動モデルによる粒子フィルタに基づく3次元体内動態推定法の追従性向上の試み

川浪佑介, 市地慶, 田辺隼騎, 奥畑孝太郎, 張彰, Zeng Yuwen, 本間経康

計測自動制御学会東北支部 60周年記念研究集会　2024年12月17日
Reliability investigation of deep learning when using imbalanced chest X-ray image sources for COVID-19 detection.

Zhang, Z, Zhang, X, Homma, N

2023 Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers　2023年
How intra-source imbalance affects deep learning performance for COVID- 19 diagnosis using chest X-ray images

Zhang Zhang, Xiaoyong Zhang, Kei Ichiji, Ivo Bukovsky, Noriyasu Homma

International Symposium on Human Welfare Engineering for Smart-Aging　2022年
Quantitative Evaluation of Explainable AI for COVID-19 Detection in Chest X-ray Images.

Zhang, Z, Zhang, X, Wang, J, Zeng, Y, Ichiji, K, Homma, N

FAN 2021 Online　2021年
顕著性マップを用いた死後CTの溺死鑑別における深層学習モデルの性能向上の試み

水野泰平, 張曉勇, 張彰, 市地慶, 杉田典大, 本間経康

第31回インテリジェント・システム・シンポジウム

︎全件表示 ︎最初の5件までを表示