Multimodal Emotion Recognition on RAVDESS Dataset Using Transfer Learning

Benchmark Model Rank Results
emotion-recognition-on-ravdessLogistic Regression on posteriors of the CNN-14&biLSTM-GuidedST–Accuracy: 80.08%
facial-emotion-recognition-on-ravdessGuided-ST and bi-LSTM with attention–Accuracy: 57.08%
speech-emotion-recognition-on-ravdessAlexNet (FineTuning)–Accuracy: 61.67%
speech-emotion-recognition-on-ravdessCNN-14 (Fine-Tuning)–Accuracy: 76.58%