• 设为首页
  • 点击收藏
  • 手机版
    手机扫一扫访问
    迪恩网络手机版
  • 关注官方公众号
    微信扫一扫关注
    公众号

parthe/Speaker-Diarization-toolkit-MATLAB: An end-to-end MATLAB toolkit for comp ...

原作者: [db:作者] 来自: 网络 收藏 邀请

开源软件名称(OpenSource Name):

parthe/Speaker-Diarization-toolkit-MATLAB

开源软件地址(OpenSource Url):

https://github.com/parthe/Speaker-Diarization-toolkit-MATLAB

开源编程语言(OpenSource Language):

MATLAB 100.0%

开源软件介绍(OpenSource Introduction):

Matlab-speaker-diarization-toolkit

An end-to-end MATLAB toolkit for completely unsupervised Speaker Diarization using state-of-the-art algorithms.

About the System

The system is useful for researchers starting their work in Speaker Diarization esp. for segmentation of broadcast news. The speech activity detector (SAD) and speaker segmentation blocks are completely unsupervised and do not require external training data. The speaker clustering is equipped with i-vector based ILP clustering which is the current state-of-the-art.

The sub-systems of the toolkit can also be plugged into other projects but have not been optimized for it. Eg: Time-series change detection, speech activity detection, Speaker recognition, Hard clustering, Soft Clustering, k-centres clustering

How to run

A few other open-source toolkits have been used. To run the system:

  1. Download the source code of this toolkit

  2. Download the dependencies by clicking the links next to names of toolkits mentioned below. They are all MATLAB codes and only need to be added to the MATLAB path. It runs smoothly on MATLAB 2013+

    i. Voicebox http://www.ee.ic.ac.uk/hp/staff/dmb/voicebox/voicebox.zip

    ii. A speech/music discriminator based on RMS and zero-crossings http://www.mathworks.com/matlabcentral/fileexchange/submissions/42092/v/7/download/zip

    iii. MSR Identity toolbox http://ftp.research.microsoft.com/downloads/2476c44a-1f63-4fe0-b805-8c2de395bb2c/MSR%20Identity%20Toolkit%20v1.0.zip?

    iv. Cluster Toolbox (Purdue) https://engineering.purdue.edu/~bouman/software/cluster/cluster-matlab/gaussmix-v1.2.zip

  3. Cite the following Thesis in your work http://www.slideshare.net/ParthePandit/parthepandit10d070009ddpthesis-53767213

This system was developped by Parthe Pandit as part of his Masters thesis. It is a Speaker Diarization system designed for segmention Broadcast News Audios. For details, please have a look at the following thesis. http://www.slideshare.net/ParthePandit/parthepandit10d070009ddpthesis-53767213




鲜花

握手

雷人

路过

鸡蛋
该文章已有0人参与评论

请发表评论

全部评论

专题导读
热门推荐
阅读排行榜

扫描微信二维码

查看手机版网站

随时了解更新最新资讯

139-2527-9053

在线客服(服务时间 9:00~18:00)

在线QQ客服
地址:深圳市南山区西丽大学城创智工业园
电邮:jeky_zhao#qq.com
移动电话:139-2527-9053

Powered by 互联科技 X3.4© 2001-2213 极客世界.|Sitemap