search

Sehaba95 / Emotions-recognition-from-audio-signal

Emotions recognition from audio signal using OpenSmile, PCA and set of classifiers from Scikit-learn library
45 stars 5 forks source link

Need help for emotion recognition #1

Open ScreamShiv opened 6 years ago

ScreamShiv commented 6 years ago

Please give a modification for extracting emotions from an unknown audio file.

Sehaba95 commented 6 years ago

How to extract features using OpenSmile ?

ScreamShiv commented 6 years ago

please provide a description of emotion extraction using this module of yours

ScreamShiv commented 6 years ago

please give me a method to extract emotions from unknown audio file.

On Thu, Jun 28, 2018 at 3:17 PM, Sehaba Amine notifications@github.com wrote:

How to extract features using OpenSmile ?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Sehaba95/Emotions-recognition-from-audio-signal/issues/1#issuecomment-400978515, or mute the thread https://github.com/notifications/unsubscribe-auth/AjYIMdUNzUSjnXe1cdsFvkXSfZ2sUtTdks5uBKYWgaJpZM4U7DKp .

Sehaba95 commented 6 years ago

You have to install OpenSmile and below there is a book and a tutorial on LinkedIn on how to install it: Book :openSMILE (the Munich open Speech and Music Interpretation by Large Space Extraction toolkit) Tutorial :Opensmile Toolkit Quick Setup

Good luck

Hooi0103 commented 4 years ago

Hello, may i know is the input .csv file created by openSmile? Or you have rewrite the MFCC feature extract by openSmile into .csv file? Or you create the Custom Configure File by yourself that can extract the all MFCC feature in SAVEE into .csv file? Because i just find out can extract MFCC features of one .wav file into HTK parameter file format only. Or actually have any ways to do it, but i did't find out yet? Can give my suggestion how to obtain the input .csv file that you get? Thank You very much

Sehaba95 commented 4 years ago

Hello @Hooi0103,

For the .csv file, I have created a custom shell and Python scripts that create it. For the Shell script: it loops over all the audio file in the SAVEE dataset and for each file it execute the OpenSmile command line to create the MFCC file and then, execute the Python script and give as input the name of the created file. In this script, it open the created file, and go to a specific line and copy the MFCC feature and paste it into the .csv file.

I wish that the response was clear.

Hooi0103 commented 4 years ago

Thank you, @Sehaba95 . Can i know you do it in Unix, Linux or Window?

Sehaba95 commented 4 years ago

I use Ubuntu 18.04