Emotion Recognition Research
Central South University
Research Assistant in Prof. Zhenlao-Liu's Lab
Time: Mar.2014 - Sept.2015
Our whole research divided to two parts: speech team and facial team. I was participated speech team.
We used CASIA Chinese emotion corpus recorded by the Institute of Automation, Chinese Academy of Sciences
Extracted speech emotion features using Praat and Matlab
Input features into the correlation analysis which is used to gain the lowest redundancy features that selected by Fisher criterion
Got recognition rate based on SVM, which is about 2% higher compared to the ones without correlation analysis
The left column contains speech emotion features without correlation analysis. The right column contains speech emotion features after correlation analysis selection.