Massive Sensing

[Japanese], Research Index, CVIM Lab, www.kameda-lab.org 2010/08/30

Abstract

As sensors are getting smaller and cheaper, it will be pretty easy to install a massive number of sensors everywhere in the very near future. However, researchers mainly have been focused on developping the theories that maximizes the efficiency of relatively limited number of sensors. On the contrary, we are going to formalize a new theory that utilizes the advantage of the massiveness of the number of sensors and their extra-long observation (training) period.

The word "Massive Sensing" is a new phrase we have been developping. It shows a research framework that premises a large number of sensors that are embedded in our daily environment and run for a very long period.


Pedestrian Localization by Human-borne Camera in Both Indoor and Outdoor Sceces

ICPR2010 We propose a localization method in town by only a single image taken by a human-borne camera. It could be applicable both in indoor scenes and outdoor scenes, it never needs other sensory devices. The method is first developped to support visually impaired people.

Intelligent Video Processing and Human Behavior Understanding by networked sensors

event recognition A method for human behavior understanding is proposed. We use imaging sensors and microphone-arrays.
Unlike "3D scene understanding and its filming", the knowledge of human behavior/event will be obtained by the system automatically.

Monitoring Assistance for Daily Activities on Large Number of Cameras

Monitoring Assistance for Daily Activities on Large Number of Cameras (Click the left image to view a large picture.)

Similarity-Based Activity Classification on Multi-Modal Sensors

Similarity-Based Activity Classification on Multi-Modal Sensors (Click the left image to view a large picture.)

Self-Calibration of Cameras without Referring Points

Self-Calibration of Cameras without Referring Points (Click the left image to view a large picture.)

Sound Source Localization on Un-Calibrated Microphones

Sound Source Localization on Un-Calibrated Microphones (Click the left image to view a large picture.)

Publication

Please check related papers/artcles in the list.

Join us!

We welcome researchers/students who are interested in this topic and think to join our research projects. Discussions and brainstormings are also welcomed. [2008/02/29]

Members

As of 2008/02/29.

Leader :
Yoshinari Kameda

Researchers :
Yoshinari Kameda(2004-)
Yuichi Ohta(2004-)

Cooperative researcher :
Itaru Kitahara (2005-)

Students :
Takashi Nishizaki(2004-), Takashi Hattori(2004-)
Koji Kanari(2005-)
Tomoyuki Kobayashi(2005-)
Shunichiro Furuhata(2006-)
Tsukasa Sumiya(2007-)


kameda[at]iit.tsukuba.ac.jp