| Item type |
紀要論文 / Departmental Bulletin Paper(1) |
| 公開日 |
2021-05-10 |
| タイトル |
|
|
タイトル |
画像と音声情報を用いたうつ病CADの高性能化 |
|
言語 |
ja |
| タイトル |
|
|
タイトル |
Developing High Performance CAD for Depression by Employing Image and Voice Information |
|
言語 |
en |
| 言語 |
|
|
言語 |
jpn |
| キーワード |
|
|
言語 |
en |
|
主題Scheme |
Other |
|
主題 |
Depression |
| キーワード |
|
|
言語 |
en |
|
主題Scheme |
Other |
|
主題 |
Voice information |
| キーワード |
|
|
言語 |
en |
|
主題Scheme |
Other |
|
主題 |
Machine learning |
| キーワード |
|
|
言語 |
en |
|
主題Scheme |
Other |
|
主題 |
SVM |
| キーワード |
|
|
言語 |
en |
|
主題Scheme |
Other |
|
主題 |
Feature selection |
| 資源タイプ |
|
|
資源タイプ識別子 |
http://purl.org/coar/resource_type/c_6501 |
|
資源タイプ |
departmental bulletin paper |
| ID登録 |
|
|
ID登録 |
10.34411/00032042 |
|
ID登録タイプ |
JaLC |
| 著者 |
和田, 昇太
牧, 優太
安倍, 和弥
武尾, 英哉
永井, 優一
Wada, Shota
Maki, Yuta
Abe, Kazuya
Takeo, Hideya
Nagai, Yuichi
|
| 抄録 |
|
|
内容記述タイプ |
Abstract |
|
内容記述 |
The number of depressed people is on the rise in Japan. In 2011, depression was added to the four major diseases, making it one of the five major diseases. This suggests the need for reliable detection of depressed people. Currently, the diagnosis of depression is mainly based on the interview. However, this lacks objective indicators, and some have questioned the accuracy of the diagnostic results. In order to solve this problem, we are developing an AI-based system that calculates the degree of certainty that a patient is depressed by focusing on facial images and voice information. The purpose of this study is to use this as an objective indicator to assist in the diagnosis of depression. This paper describes the construction of a voice discriminator that constitutes the system. A total of 42 features were created based on the patient's speech and the characteristics of the depressed person. The SVM was used to create and evaluate the discriminator, and the accuracy was 90% for the female data. However, the accuracy for male data was about 55%. Hence, these features were selected by 18 feature selection methods. The discriminator was made and evaluated again, and the accuracy was 80%, even for male data. |
|
言語 |
en |
| 書誌情報 |
神奈川工科大学研究報告.B,理工学編
巻 45,
p. 9-15,
発行日 2021-03-01
|
| 出版者 |
|
|
出版者 |
神奈川工科大学 |
| ISSN |
|
|
収録物識別子タイプ |
PISSN |
|
収録物識別子 |
21882878 |
| 書誌レコードID |
|
|
収録物識別子タイプ |
NCID |
|
収録物識別子 |
AA12669200 |
| フォーマット |
|
|
内容記述タイプ |
Other |
|
内容記述 |
application/pdf |
| 著者版フラグ |
|
|
出版タイプ |
VoR |
|
出版タイプResource |
http://purl.org/coar/version/c_970fb48d4fbd8a85 |
kkb-045-002.pdf (1.1 MB)
