Voice-Based Age, Gender Recognition using LSTM Architecture

Please press ⭐ button and/or cite papers if you feel helpful.

The Problem

Develop an application which facilitates:

• Gender recognition of a person based on a voice recording.
• Age group identification of a person based on a voice recording.

The dataset

For this project we've used

Download Mozilla Common Voice Dataset

, namely, the English dataset for gender classification and age classification. And the English, French and German datasets for language classification. The gender dataset contains 230,000 samples, labeled as male, female or other. The age dataset contains 250,000 samples labeled from teens to eighties, in tens of years increments. The language dataset is self-explanatory.

Usage

Preprocessing

Update soon

Training Process

Update soon

Testing and visualize evaluate

Update soon

Contact

For any information, please contact the main author:

Nhut Minh Nguyen at FPT University, Vietnam

Email: [email protected]

GitHub: https://github.com/nhut-ngnn