本文へジャンプ

Laboratory

>>Lab. list

Artificial Intelligence

Towards Man-Machine Symbiotic Systems

Our research objective is to realize intelligent technologies for understanding a variety of complicated situations in the real world. More specifically, the goal of our research is to develop fundamental technologies required not only for the virtual networked society but also for a next-generation ICT ubiquitous society where human beings, physically-embodied robots, and intelligent appliances live symbiotically while they interact each other. In particular, we put our focus upon the Visual Information Processing essential for real-world understanding and the Embodied Information Processing indispensable in physical interaction.


Staff Prof. KIDODE, Masatsugu
Assoc.Prof. UKITA, Norimichi
Assist.Prof. HABE, Hitoshi and MATSUBARA, Takamitsu
Laboratory's HP
Prof. KIDODE
Prof. KIDODE

Research Area

Visual Information Processing (Integrated Vision for Understanding the Real World)

A human recognizes a situation in the real world mostly from visual information. Our lab studies to develop sophisticated visual-perception functions in a computer system by integrating analytical results acquired from multiple cameras. In particular, we focus on human detection, tracking, 3D reconstruction, and pose/motion estimation, which are essential for understanding dynamic situations in the real world. For achieving these tasks, we are interested in wide-area observation and camera control/selection for obtaining detailed information of observed people. We are also interested in human modeling from observed images, such as modeling physical and social capabilities.

Embodied Information Processing (Networked-Machine Cooperation)

The objective is to realize robotic systems that can support our daily activities (e.g., providing useful information and boosting up and carrying out for human's jobs) by understanding intentions/behaviors of human beings. In particular, we study how to ingenerate advanced features from cooperative behaviors among embodied robots and machines; for example, complementary behaviors among heterogeneous robots, functional augmentation by cooperative behaviors among homogeneous robots (e.g., power/speed augmentation), and robot motion learning by imitating human motion.

左:人体・着衣の姿勢・形状分析 、右:人物の動作軌跡の解析
左:人体・着衣の姿勢・形状分析 、右:人物の動作軌跡の解析
左:運動スキル解析 、右:卓球ロボット
左:運動スキル、右:卓球ロボット

Research Facilities

  • マルチカメラシステム
  • 高機能双碗ロボットアーム・ハンドシステムを始めとする各種ロボットなど

Feature of the Laboratory

Visual and embodied functionalities are designed and developed in a computer system. These functionalities are then employed for modeling and analyzing human motion and behavior.

Such video analysis/synthesis and robot development have been fascinating many students, including Ph.D candidates and those from companies and overseas.

We are offering motivated students a nice environment for research and education. For example, in recent years, some students have opportunities to study at foreign research institutes, where faculty members stay for cooperative research.

3次元ビデオ・身体運動の撮影スタジオ・システム
3次元ビデオ・身体運動の撮影スタジオ・システム

Joint Researches and Social Activities

We are promoting cooperative research activities with Keihanna Joint Graduate School and other universities, research institutes, and companies in Japan and overseas.

Recent Theses and Dissertations

Ph.D. Dissertations
Retno Supriyanti, “Cataract Screening Techniques under Limited Health Facilities”
満上育久, “Multiple steerable projector system for geometrically correct displays in a wide environment” Summary PDFThesis PDF
山本大介(社会人学生), “Home robot interface innovated by familiar behaviors” Summary PDFThesis PDF
Sofiane Yous, “Active Camera Control for High-resolution Imaging” Summary PDFThesis PDF
Master's Theses
中河秀仁, “人物の移動軌跡と画像の色情報を用いた人物存在事前確率の効率的な獲得” Thesis PDF
平井迪郎, “姿勢推定のための確率的な形状変化の動的モデルと視体積制約に基づく形状追跡” Thesis PDF