CNN-Transformer hybrid reaches 98.1% accuracy on Arabic SER using EYASE and BAVED datasets, outperforming CNN-LSTM and fine-tuned wav2vec 2.0.
Survey on speech emotion recognition: Features, clas- sification schemes, and databases,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Towards Robust Arabic Speech Emotion Recognition with Deep Learning
CNN-Transformer hybrid reaches 98.1% accuracy on Arabic SER using EYASE and BAVED datasets, outperforming CNN-LSTM and fine-tuned wav2vec 2.0.