| Project/Area Number |
13650430
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
情報通信工学
|
|---|
| Research Institution | MEIJI UNIVERSITY |
|---|
Principal Investigator |
ARAKAWA Kaoru MEIJI UNIVERSITY, SCHOOL OF SCIENCE AND TECHNOLOGY, PROFESSOR, 理工学部, 教授 (30183734)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
ITOH Setsuko DOSHISHA WOMEN'S COLLEGE, DEPT. OF HUMAN LIFE STUDIES 9., PROFESSOR, 生活科学部, 教授 (50144358)
|
|---|
| Project Period (FY) |
2001 – 2002
|
| Project Status |
Completed (Fiscal Year 2002)
|
| Budget Amount *help |
¥4,100,000 (Direct Cost: ¥4,100,000)
Fiscal Year 2002: ¥1,600,000 (Direct Cost: ¥1,600,000)
Fiscal Year 2001: ¥2,500,000 (Direct Cost: ¥2,500,000)
|
|---|
| Keywords | baby cries / speech analysis / babies and infants / signal processing / childcare / state estimation / nonlinear filter / noise reduction / 音声解析 / 泣き声 / 非線形ディジタルフィルタ / ルールベースシステム / 意味情報解析 / サウンドスペクトログラム |
|---|
| Research Abstract |
A way of computer analyses is proposed for crying babies' voice in order to recognize their mental status.
Since babies do not have language, they just cry when they are in some uncomfortable situation. Mothers try to avoid the cause for the uncomfortableness, but it is sometimes difficult to recognize why the babies are crying by just observing them. In this case, mothers cannot stop the babies crying and then get into great trouble. Such a problem in taking care of babies can be solved, if the mental situation of babies is recognized automatically by analyzing some signs from them.
In this research, a method for recognizing the mental status of crying babies is proposed by analyzing their voice in the frequency domain. Especially, a signal processing and rule-base system is developed to classify the crying voice when they are hungry and that when they are sleepy. Applying the system to some real babies voice, we showed that about 80% of the data were classified correctly by this system, although they cannot be classified correctly by human audition. The baby cries when they had pain are also examined. Moreover, a new type of nonlinear digital filter is proposed to reduce the noise from environment which is added to speech signal. The high performance of this filter is also verified for actual speech signals. This filter is effective as a pre-processing system of the baby cries classifier.
|
