TESS speech emotion recognition with log-mel CNNs, speaker-independent train (OAF) / test (YAF), SpecAugment-style augmentation, Grad-CAM explainability, and a small Whisper+TF‑IDF text baseline showing emotion isn’t in transcripts alone.
machine-learning computer-vision deep-learning tensorflow grad-cam artificial-intelligence tf-idf convolutional-neural-networks audio-classification librosa data-augmentation ccn speech-emotion-recognition explainable-ai mel-spectrogram tess-dataset speaker-independent-evaluation
-
Updated
May 10, 2026 - Jupyter Notebook