ISCA Archive Interspeech 2013
ISCA Archive Interspeech 2013

Voice pathology detection and classification using MPEG-7 audio low-level features

Ghulam Muhammad, Moutasem Melhem

In this paper, a new pathological voice detection and pathology classification method based on MPEG-7 audio low-level features is proposed. MPEG-7 features are originally used for multimedia indexing, which includes both video and audio. Indexing is related to event detection, and as pathological voice is a separate event than normal voice, we show that MPEG-7 audio low-level features can do very well in detecting pathological voices, as well as classifying the pathologies. The experiments are done on a subset of sustained vowel (namely, "AH") recordings from healthy and voice pathological subjects, from the MEEI database. For classification, support vector machine (SVM) with 10-fold cross-validation is applied. The proposed method with MPEG-7 audio features and SVM classification is evaluated on voice pathology detection, as well as pathology classification. The experiment results show that the proposed method outperforms some recent methods in the literature both in detection and in classification. The proposed method is able to achieve an accuracy of 99.994 } 0.0105% for detecting pathological voices and an accuracy of 100% for binary pathologies classifying.


doi: 10.21437/Interspeech.2013-321

Cite as: Muhammad, G., Melhem, M. (2013) Voice pathology detection and classification using MPEG-7 audio low-level features. Proc. Interspeech 2013, 3627-3631, doi: 10.21437/Interspeech.2013-321

@inproceedings{muhammad13_interspeech,
  author={Ghulam Muhammad and Moutasem Melhem},
  title={{Voice pathology detection and classification using MPEG-7 audio low-level features}},
  year=2013,
  booktitle={Proc. Interspeech 2013},
  pages={3627--3631},
  doi={10.21437/Interspeech.2013-321}
}