Mitsuharu Matsumoto Laboratory  

Search WWW Search
Japanese | English
About us
Research projects
->Image processing
->Acoustical signal processing
->Intelligent robotics
->Interaction design
->Artificial intelligence
For students
Access map
Inside mm-labo
Subjective computing / Artificial intelligence

Many systems such as noise reduction system and feature extraction system in acoustical signal processing and image processing require parameter settings to obtain adequate outputs.

Although many studies aim to solve these types of problems by using the objective relation between signal and noise, it is also important to take into account subjective prospects of humans.

Mitsuharu Matsumoto Laboratory progresses our research from objective and subjective prospects such as:

- Development of light and noise robust human detection system
- Subjective parameter setting using facial recognition system
- Parameter optimization using speech recognition system
- Parameter optimization based on signal-noise decorrelation

We would like to study subjective computing, which can handle not only objective perspective but also subjective perspective in perceptual information processing.


Noise and light robust human detection system

Although human detection system becomes practical in various scenes such as camera and television system, it is still difficult to detect human in noisy and light variant environments. We are studying noise and light robust human detection system.

Parameter setting using face recognition system

When the filter output is not the original image itself but feature images such as edge images and portrait images, it is difficult to use objective criterion for parameter setting. We are studying parameter setting using face recognition system to handle parameter setting of edge images and portrait images.

Parameter setting using speech recognition system

We aim to study parameter settting using speech recognition system. We also would like to apply our approach to self-calibration of humanoid robot.

Parameter setting based on signal-noise decorrelation

This research aims to realize parameter setting based on signal-noise decorrelation.

(C) Copyright 2009 - Mitsuharu Matsumoto, The university of electro-communications.