專利授權區國立清華大學國際產學營運總中心 Operations Center for Industry Collaboration

搜尋專利授權區

關鍵字

» 新增關鍵字

選單

專利授權區

專利授權區
專利名稱(中)	語音活動檢測的計算機實施方法、系統與電腦程式產品
專利名稱(英)	COMPUTER-IMPLEMENTED METHOD, SYSTEM AND COMPUTER PROGRAM PRODUCT FOR VOICE ACTIVITY DETECTION
專利家族	中華民國：I922079
專利權人	國立清華大學 100.00%
發明人	李祈均,吳柏澄
技術領域	資訊工程,電子電機

專利摘要(中)
本發明提出一種語音活動檢測的計算機實施方法。所述計算機實施方法基於多個特徵提取器分別提取輸入話語的多個第一特徵，並且基於這些第一特徵，通過預訓練分類器判斷輸入話語是否對應於目標對象。所述多個特徵提取器是分別由對應不同的多個情境的多個訓練集來訓練。此外，使用此方法的系統和電腦程式產品亦被提出。

專利摘要(中)

本發明提出一種語音活動檢測的計算機實施方法。所述計算機實施方法基於多個特徵提取器分別提取輸入話語的多個第一特徵，並且基於這些第一特徵，通過預訓練分類器判斷輸入話語是否對應於目標對象。所述多個特徵提取器是分別由對應不同的多個情境的多個訓練集來訓練。此外，使用此方法的系統和電腦程式產品亦被提出。

專利摘要(英)
A computer-implemented method for voice activity detection (VAD) is provided. According to this computer-implemented method, multiple first features of an input utterance are extracted based on multiple feature extractors respectively, and whether the input utterance corresponds to a target object is determined through a pre-trained classifier based on the multiple first features. The multiple feature extractors are respectively trained by multiple training sets corresponding to multiple different scenarios. In addition, a system and a computer program product using this method are also provided.

專利摘要(英)

A computer-implemented method for voice activity detection (VAD) is provided. According to this computer-implemented method, multiple first features of an input utterance are extracted based on multiple feature extractors respectively, and whether the input utterance corresponds to a target object is determined through a pre-trained classifier based on the multiple first features. The multiple feature extractors are respectively trained by multiple training sets corresponding to multiple different scenarios. In addition, a system and a computer program product using this method are also provided.

聯絡資訊
承辦人姓名	李曉琪
承辦人電話	03-5715131 #31061
承辦人Email	hsiaochi@mx.nthu.edu.tw