AI RESEARCH
Hybrid CNN-Transformer Architecture for Arabic Speech Emotion Recognition
arXiv CS.CL
•
ArXi:2604.07357v1 Announce Type: new Recognizing emotions from speech using machine learning has become an active research area due to its importance in building human-centered applications. However, while many studies have been conducted in English, German, and other European and Asian languages, research in Arabic remains scarce because of the limited availability of annotated datasets. In this paper, we present an Arabic Speech Emotion Recognition (SER) system based on a hybrid CNN-Transformer architecture.