Extracting deep bottleneck features for visual speech recognition

Research output: Chapter in Book/Conference paperConference paper

9 Citations (Scopus)

Abstract

© 2015 IEEE. Motivated by the recent progresses in the use of deep learning techniques for acoustic speech recognition, we present in this paper a visual deep bottleneck feature (DBNF) learning scheme using a stacked auto-encoder combined with other techniques. Experimental results show that our proposed deep feature learning scheme yields approximately 24% relative improvement for visual speech accuracy. To the best of our knowledge, this is the first study which uses deep bottleneck feature on visual speech recognition. Our work firstly shows that the deep bottleneck visual feature is able to achieve a significant accuracy improvement on visual speech recognition.
Original languageEnglish
Title of host publicationICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages1518-1522
Volume2015-August
ISBN (Print)9781467369978
DOIs
Publication statusPublished - 2015
EventExtracting deep bottleneck features for visual speech recognition - South Brisbane, Queensland
Duration: 1 Jan 2015 → …

Conference

ConferenceExtracting deep bottleneck features for visual speech recognition
Period1/01/15 → …

    Fingerprint

Cite this

Sui, C., Togneri, R., & Bennamoun, M. (2015). Extracting deep bottleneck features for visual speech recognition. In ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings (Vol. 2015-August, pp. 1518-1522). IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ICASSP.2015.7178224