透過您的圖書館登入
IP:18.216.34.146
  • 期刊

笑聲偵測技術文獻回顧

A Review on Laughter Detection Technology

摘要


在許多的人機互動的應用中,使用者情緒辨識可以用來取得使用者額外的資訊並適時導引系統做正確的回應,使系統更人性化。而偵測「附屬語言活動」常被用於使用者情感狀態的分類,其中一個最明顯的「附屬語言事件」就是笑聲,所以本論文將著眼於自動笑聲偵測技術做一個簡短的文獻回顧,介紹笑聲的基本聲學特性、常用的實驗資料庫、特徵參數與笑聲分類器的建立。我們期待笑聲偵測技術可以在許多應用場景,提供更有趣且多元化的人機互動機制。

並列摘要


In many human-machine interaction applications, user's emotion can be recognized to reveal the additional user's information which can improve the system's response and make the system to be more user-friendly. Detecting the ”paralinguistic events” is a way to classify the user's emotion state. One of the most obvious ”paralinguistic events” is laughter. The paper provides a brief review focusing on the studies of automatic laughter detection technology. The topics about the acoustic properties of laughter, the famous experimental database, extracted features and the laughter classifiers are discussed. We expect the laughter detection technology can be applied in many application scenarios that provide more interesting human-machine interaction mechanism.

延伸閱讀


國際替代計量