mightbesimon / speech2emotion

classify emotion from speech signal

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Digital Signal Processing

ELECTENG 733
University of Auckland
Simon Shan

Emotion recognition - Practical Implementation Assignment 2
May 2020

  • angry
  • excited
  • happy
  • sad

data preperation and feature extraction done for all 60 audio signals (together), but traning (i.e. the mle function) is only done on the first 48, classification is done on the last 12 (using the extracted features)

Background

Human-Computer Interaction applications like healthcare robots, talking aids and devices that interact with humans socially require robust emotion recognition systems. Using this emotion recognition capability, Human-Computer Interactive technology can understand human emotions and respond accordingly. Emotion recongition can be done from facial expressions, speech and gestures. In this assignment, your task is to build an emotion recognition system with speech signal as the input. With the power of signal processing, statistical analysis and machine learning, robust emotion recognition systems have been developed by researchers. However, human emotions are complex, and the features of the speech signal vary depending on the age, gender, accent-type and language of the speaker. Hence, the task becomes complicated. Good knowledge about the speech signal, signal processing techniques and decision making are key stages that can help to produce better speech recognisers. In this assignment, you will go through the process of developing an emotion recogniser from speech signal, implementing majority of the concepts you learnt during Part 1 of ELECTENG 733. It will consist of a Training stage and a Testing stage.

assignment brief
https://www.isca-speech.org/archive/Interspeech_2018/pdfs/1349.pdf

Requirements

MATLAB (download)

run main.m

Explaination

graph

Speech Features

  • short_time_energy.m
  • zero_crossing_rate.m
  • pitch.m
  • spectral_energy.m

Authors

  • simon - buy my merch hire me pls - mightbesimon @mightbesimon

License

MIT

Acknowledgments

  • these are just my sample codes, if you misuse them its not my problem
  • not actually in ELECTENG 733, just felt like doing the project for fun

About

classify emotion from speech signal


Languages

Language:MATLAB 100.0%