/EmotionRecognition_2Dcnn-lstm

The code ruproduced the emotion recognition model, 2D CNN LSTM networks, which based on <Speechemotionrecognitionusingdeep1D&2DCNNLSTMnetworks>, Jianfeng Zhao et.

Primary LanguagePythonBSD 3-Clause "New" or "Revised" LicenseBSD-3-Clause

EmotionRecognition_2D CNN LSTM networks

Introduction

According to the nice paper,Speech emotion recognition using deep 1D & 2D CNN LSTM networks,the 2D CNN LSTM model was built by tensorflow2-keras modul.With training and testing in EmoDB, the model we built showed the closest conclusion comparead with the paper.

Requirements

The code should run the enviroment as follow list:

name version
python 3.8
numpy 1.19.2
tensorflow 2.2.0
librosa 0.8.0
scikit_learn 0.24.1

Before running the code, you sholud set up the enviroment we needed by entering the following command into the terminal:

  • pip install -r requirement.txt

and then verify the parameter of dataset path in main.py

  • __EmoDB_file_path__ = 'your_dataset_path'

and finally, running!

Dataset

You can dowload Berlin Database of Emotional Speech.

Reference

JianfengZhao,Xiao Mao,Lijiang Chen, Speech emotion recognition using deep 1D & 2D CNN LSTM networks,Biomedical Signal Processing and Control Volume 47, January 2019, Pages 312-323

Contact

Any issuse should submit directly or send email msp_baiyuhe@163.com.