Sapporo, Japan
Sapporo, Japan
Nara Institute of Science and Technology
(NAIST), Japan
Nara Medical University, Japan
Yoshinobu Sato received his B.S., M.S. and Ph.D. degrees in Information and Computer Sciences from Osaka University, Japan in 1982, 1984, and 1988, respectively. From 1988 to 1992, he worked at the NTT Human Interface Laboratories. In 1992, he joined Osaka University Medical School. From 1996 to 1997, he was a Visiting Research Fellow at the Surgical Planning Laboratory, Harvard Medical School and Brigham and Women's Hospital. In 1999, he became an Associate Professor at Osaka University Graduate School of Medicine. From 2014 to 2025, he was a Professor at the Nara Institute of Science and Technology (NAIST). Currently, he is a Professor Emeritus of NAIST and an Invited Professor at Nara Medical University.
His research interests include medical image analysis, computer-assisted surgery, and computational anatomy. Dr. Sato is a recipient of the CAOS-International Maurice E. Müller Award for Excellence in Computer Assisted Surgery. He serves as an Editorial Board member of the Medical Image Analysis journal and IEEE Transactions on Medical Imaging, and is a Fellow of the MICCAI (Medical Image Computing and Computer Assisted Intervention) Society.
Speech Title: "Next-Generation Computer-Assisted Orthopedic Surgery: AI-Driven Musculoskeletal Image Analysis and Surgical Planning"
Abstract: Recent advancements in deep learning and image analysis technologies have enabled high-precision, automated recognition of musculoskeletal structures from medical images. In CT imaging, this allows for the automated, quantitative analysis of individual muscle shape, volume, and density alongside traditional bone metrics. By automatically processing large-scale clinical CT datasets, surgical information, and patient records, this approach provides objective, quantitative support for orthopedic surgical decision-making and planning—domains that have traditionally relied on empirical evaluation.
Based on these automated recognition results, we have developed large-scale training datasets to analyze age-related anatomical variations, track disease progression, and automate surgical planning. Furthermore, we are developing AI models to estimate quantitative musculoskeletal parameters directly from plain X-ray images. This framework facilitates complementary preoperative planning and surgical indication assessment using both CT and X-ray imaging, while also enabling low-dose postoperative evaluation and long-term preventive monitoring. In this keynote lecture, we present our musculoskeletal AI analysis infrastructure and discuss its practical applications for next-generation orthopedic surgery support.
Department of Biomedical Engineering, National Cheng Kung University, Tainan, China
Dr. Tzu-Hsien Yang is an associate professor in the Department of Biomedical Engineering at National Cheng Kung University (NCKU), Tainan, China. He received his BS and PhD degrees in electrical engineering from NCKU in 2010 and 2014, respectively. His research interests encompass computational systems biology, bioinformatics, and deep learning. Dr. Yang has published extensively in systems biology and bioinformatics, focusing on developing algorithms and tools that integrate multi-omic and high-throughput data to model and analyze biological systems. Additionally, he has substantial experience in developing AI systems for clinical and medical diagnosis.
Speech Title: "Prediction of histone signals for identifying cell-type-specifically active cis-transcriptional regulatory modules"
Abstract: Cis-regulatory modules (CRMs) are genomic DNA regions bound by transcription factors. CRMs can thus control the spatial and temporal gene expression in cells. Accurate identification of cell-type-specifically active CRMs is essential for understanding transcriptional regulation across diverse biological contexts. They are also closely associated with disease, as aberrant CRM activity can disrupt cell-type-specific gene regulation and contribute to disease-related transcriptional programs. Therefore, accurate identification of cell-type-specifically active CRMs is important for understanding transcriptional regulation and disease mechanisms. It has been shown that cell-type-specifically active CRMs can be predicted from cell-type-specific histone ChIP-seq (chromatin immunoprecipitation sequencing) signals. However, generating histone ChIP-seq data for every relevant cell type is experimentally costly, particularly for rare or disease-associated cell populations. Further, no existing tool provides reliable cell-type-specific histone ChIP score prediction required for systematic CRM analysis. To address this limitation, we developed a cross-cell-type computational model that predicts cell-type-specific histone ChIP scores by integrating more accessible ATAC-seq, CAGE-seq, and RNA-seq data. Using these predicted histone ChIP scores, we facilitated a computational framework for effectively identifying cell-type-specifically active CRMs without requiring direct histone ChIP-seq experiments. This approach provides a scalable strategy for CRM annotation in insufficiently profiled cell types and advances the study of transcriptional regulation for diseases.
