Biomedical Imaging Intelligence

Research Staff

  • Associate Professor Yoshito OTAKE

    Associate Professor
    Yoshito OTAKE

E-mail otake@is.naist.jp
To the site Under construction

Research Areas

We conduct research on intelligent systems to analyze biomedical images for future medical applications. By leveraging artificial intelligence trained on various types of biomedical image big data, we aim to enhance diagnosis, treatment, and prognosis prediction, while also developing fundamental AI technologies motivated by clinical needs.

Intelligence for Image Understanding

We aim to construct a "virtual human body" within a computer by comprehensively recognizing three-dimensional biomedical images of the human body. This involves modeling anatomical knowledge through advanced methods such as deep learning, multidimensional image analysis, statistical modeling, and Bayesian inference, enabling automated understanding of complex biomedical structures.

Intelligence for Biomedical Big Data Mining

We utilize artificial intelligence to systematically organize and analyze vast amounts of biomedical imaging data accumulated from daily clinical practices at hospitals nationwide. Through this process, we extract prior anatomical knowledge, quantitatively analyze variations associated with age and gender, and predict whole-body anatomical structures from minimally invasive partial measurements.

Intelligence for Predicting Internal Body Dynamics

We analyze how muscles and bones interact within the body during movement, using minimally invasive measurement data. By predicting subject-specific musculoskeletal geometry, we achieve higher accuracy in analysis. These methods are applied to assess the effectiveness of rehabilitation and provide motion guidance aimed at injury prevention for athletes.

Intelligence for Treatment Design

We perform patient-specific biomechanical simulations on a virtual human body model and integrate statistical predictions derived from surgical databases to enable automated surgical planning. This approach scientifically captures the expertise and intuitive insights of experienced surgeons, facilitating the design of optimal surgical strategies tailored specifically to individual patients.