Early prediction of major depression in adolescents using glottal wave characteristics and Teager Energy parameters

Ooi, Kuan Ee Brian; Low, Lu-Shih Alex; Lech, Margaret; Allen, Nicholas B.

doi:10.1109/icassp.2012.6288946

Cited by 16 publications

(6 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Even though F0 is restricted to voiced speech, its performance was low compared with most other features, which indicates that F0 works better for speaker-dependent comparison. Interestingly, unvoiced speech performs better than voiced speech using log-energy and log-Teager-Energy in both low- [13,20,21], we investigate unvoiced and mixed speech as well. The best depression recognition results in this study were achieved by RMS-Teager-Energy giving 80% accuracy using mixed speech and Log-Teager-Energy, which was also 80% accurate using unvoiced speech.…”

Section: Classification Resultsmentioning

confidence: 99%

“…Recently, the automatic detection of depression using computer artificial intelligence techniques has been investigated [17,18,19,13,20]. While psychological investigations are concerned with the overall patterns of speech using statistical functionals of speech features, affective computing classification can be based on frame-by-frame low-level features extracted from speech.…”

Section: Introductionmentioning

confidence: 99%

“…HNR, jitter and shimmer features gave moderate results in [19], though more investigation is needed. Non-linear features have been investigated recently to detect depression [13,20]. These features are based on the Teager energy operator (TEO), which measure the number of harmonics produced from the non-linear air flow in the vocal tract.…”

Section: Introductionmentioning

confidence: 99%

“…These features are based on the Teager energy operator (TEO), which measure the number of harmonics produced from the non-linear air flow in the vocal tract. Originally, TEO based features were used to detect stressed speech [21], but they outperform linear features on detecting depression as well [13,20]. Multi-dimensional speech features were not used in this work, for the purpose of characterising depressed speech and equal comparison.…”

Section: Introductionmentioning

confidence: 99%

“…Not only is there very little work on the automatic detection of depression from speech in the literature, previous researchers applied different methods and features, which make the comparison of results even harder. Most depression studies focus the feature extraction on voiced speech only [18,13,20]. Most emotion recognition studies use either low-level or utterancelevel features.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Characterising depressed speech for classification

et al. 2013

View full text Add to dashboard Cite

Depression is a serious psychiatric disorder that affects mood, thoughts, and the ability to function in everyday life. This paper investigates the characteristics of depressed speech for the purpose of automatic classification by analysing the effect of different speech features on the classification results. We analysed voiced, unvoiced and mixed speech in order to gain a better understanding of depressed speech and to bridge the gap between physiological and affective computing studies. This understanding may ultimately lead to an objective affective sensing system that supports clinicians in their diagnosis and monitoring of clinical depression. The characteristics of depressed speech were statistically analysed using ANOVA and linked to their classification results using GMM and SVM. Features were extracted and classified over speech utterances of 30 clinically depressed patients against 30 controls (both gender-matched) in a speaker-independent manner. Most feature classification results were consistent with their statistical characteristics, providing a link between physiological and affective computing studies. The classification results from low-level features were slightly better than the statistical functional features, which indicates a loss of information in the latter. We found that both mixed and unvoiced speech were as useful in detecting depression as voiced speech, if not better.

show abstract

Section: Classification Resultsmentioning

confidence: 99%