Compares predicted severity distributions of synthetic speech methods against original pathological speech.
Classifier: ./outputs/severity_classifier_mms (accuracy: 97.87%)
Reference: original_speech
Reference distribution: Mild: 15.8%, Moderate: 35.1%, Normal: 41.5%, Severe: 7.6%
Ranking (by Jensen-Shannon distance, lower = closer to original)
original_speech
0.0000
1.0000
0.0000… See the full description on the dataset page:
https://huggingface.co/datasets/resproj007/severity_comparison_results.