TY - GEN
T1 - Speech-based diagnosis of autism spectrum condition by generative adversarial network representations
AU - Deng, Jun
AU - Cummins, Nicholas
AU - Schmitt, Maximilian
AU - Qian, Kun
AU - Ringeval, Fabien
AU - Schuller, Björn
N1 - Publisher Copyright:
© 2017 ACM.
PY - 2017/7/2
Y1 - 2017/7/2
N2 - Machine learning paradigms based on child vocalisations show great promise as an objective marker of developmental disorders such as Autism. In conventional detection systems, hand-crafted acoustic features are usually fed into a discriminative classifier (e. g., Support Vector Machines); however it is well known that the accuracy and robustness of such a system is limited by the size of the associated training data. This paper explores, for the first time, the use of feature representations learnt using a deep Generative Adversarial Network (GAN) for classifying children's speech affected by developmental disorders. A comparative evaluation of our proposed system with different acoustic feature sets is performed on the Child Pathological and Emotional Speech database. Key experimental results presented demonstrate that GAN based methods exhibit competitive performance with the conventional paradigms in terms of the unweighted average recall metric.
AB - Machine learning paradigms based on child vocalisations show great promise as an objective marker of developmental disorders such as Autism. In conventional detection systems, hand-crafted acoustic features are usually fed into a discriminative classifier (e. g., Support Vector Machines); however it is well known that the accuracy and robustness of such a system is limited by the size of the associated training data. This paper explores, for the first time, the use of feature representations learnt using a deep Generative Adversarial Network (GAN) for classifying children's speech affected by developmental disorders. A comparative evaluation of our proposed system with different acoustic feature sets is performed on the Child Pathological and Emotional Speech database. Key experimental results presented demonstrate that GAN based methods exhibit competitive performance with the conventional paradigms in terms of the unweighted average recall metric.
KW - Autism spectrum condition
KW - Automatic diagnosis
KW - Generative adversarial networks
KW - Representation learning
UR - http://www.scopus.com/inward/record.url?scp=85025426583&partnerID=8YFLogxK
U2 - 10.1145/3079452.3079492
DO - 10.1145/3079452.3079492
M3 - Conference contribution
AN - SCOPUS:85025426583
T3 - ACM International Conference Proceeding Series
SP - 53
EP - 57
BT - DH 2017 - Proceedings of the 2017 International Conference on Digital Health
PB - Association for Computing Machinery
T2 - 7th International Conference on Digital Health, DH 2017
Y2 - 2 July 2017 through 5 July 2017
ER -