Electronics and Information Engineering Laboratories

Intelligent Information Media Laboratory

Basic Information

Staff

Professor Norimichi Ukita

Outline

Intelligent information systems by multimedia processing such as computer vision, pattern recognition, and machine learning

We develop fundamental technologies for man-machine symbiotic Intelligent Information Processing (IIP) systems. These fundamental technologies—which include cooperative distributed architectures and sensor networks, multimedia sensing and recognition (e.g., computer vision and pattern recognition), and big data analysis using machine learning—are used to develop cyber-physical systems that incorporate artificial intelligence, big data, and computational resources for building IIP with international collaborators. In what follows, I will describe six of the key areas of my research experience.

Research Themes

Computer Vision, Pattern Recognition, Machine Learning, Computer Graphics, Human Sensing and Modeling, Robotics

Website

http://www.toyota-ti.ac.jp/Lab/Denshi/iim/ukita/

Activities

(1) Real-time real-world sensing using network-connected cooperative distributed active vision cameras and its application to multi-people tracking and high-precision human shape and motion analysis

reid.png
kimono_pose.png
shape_carving_graph_cuts.jpg
 
(2) Human attributes estimation in surveillance videos
gaze.png
 
(3) Human pose and activity recognition using very large samples
action_pose.jpg
 
(4) Human behavior recognition, condition evaluation, and intervention using multimodal sensors and interactions
gait.png

(5) Long-term physiological data modeling and its applications to robotics and QoL support


(6) Domain-specific modeling of high-dimensional temporal data and its application to personally-optimized physical exercise and rehabilitation


 

page up