About this document

GANs for Speech Emotion Data Augmentation by aladinthemmagic is a document available to read on EtoBox.

This document presents a study on addressing data imbalance in Speech Emotion Recognition (SER) by utilizing Generative Adversarial Networks (GANs) for data augmentation. The authors propose a modified GAN architecture to generate synthetic spectrograms for underrepresented emotional classes, demonstrating significant performance improvements on two datasets, IEMOCAP and FEEL-25k. The methodology includes autoencoder training, GAN initialization, and fine-tuning, with results indicating that the GAN-based a

Author

aladinthemmagic

Language

EN