• Login
    View Item 
    •   Home
    • Graduate Theses, Dissertations and Projects
    • Queen's Graduate Theses and Dissertations
    • View Item
    •   Home
    • Graduate Theses, Dissertations and Projects
    • Queen's Graduate Theses and Dissertations
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Toward Smart Classrooms: Automated Detection of Speech Analytics and Disfluency with Deep Learning

    Thumbnail
    View/Open
    Thesis document (1.475Mb)
    Author
    Kourkounakis, Tedd
    Metadata
    Show full item record
    Abstract
    Strong presentation skills are valuable and sought-after in workplace and classroom environments alike. Though they are necessary for day-to-day life, there is a lack of effective resources to help improve upon these abilities. This thesis presents a speech assessment tool capable of providing feedback on different metrics of speech quality during a presentation. The information provided by this application has the ability to identify obstacles in one's vocal performance and allows users to further develop these skills. Of the possible improvements to vocal presentations, disfluencies and stutters in particular remain one of the most common and prominent factors of someone's demonstration. Millions of people are affected by stuttering and other speech disfluencies, with the majority of the world having experienced mild stutters while communicating under stressful conditions. While there has been much research in the field of automatic speech recognition and language models, there lacks the sufficient body of work when it comes to disfluency detection and recognition. To this end, we propose an end-to-end deep neural network, FluentNet, capable of detecting a number of different disfluency types. FluentNet consists of a Squeeze-and-Excitation residual convolutional neural network which facilitates the learning of strong spectral frame-level representations, followed by a set of bidirectional long short-term memory layers that aid in learning effective temporal relationships. Lastly, FluentNet uses an attention mechanism to focus on the important parts of speech to obtain a better performance. We perform a number of different experiments, comparisons, and ablation studies to evaluate our model. Our model achieves state-of-the-art results by outperforming other solutions in the field on the publicly available UCLASS dataset. Additionally, we present LibriStutter: a disfluency dataset based on the public LibriSpeech dataset with synthesized stutters. We also evaluate FluentNet on this dataset, showing the strong performance of our model versus a number of benchmark techniques.
    URI for this record
    http://hdl.handle.net/1974/28165
    Collections
    • Department of Electrical and Computer Engineering Graduate Theses
    • Queen's Graduate Theses and Dissertations
    Request an alternative format
    If you require this document in an alternate, accessible format, please contact the Queen's Adaptive Technology Centre

    DSpace software copyright © 2002-2015  DuraSpace
    Contact Us
    Theme by 
    Atmire NV
     

     

    Browse

    All of QSpaceCommunities & CollectionsPublished DatesAuthorsTitlesSubjectsTypesThis CollectionPublished DatesAuthorsTitlesSubjectsTypes

    My Account

    LoginRegister

    Statistics

    View Usage StatisticsView Google Analytics Statistics

    DSpace software copyright © 2002-2015  DuraSpace
    Contact Us
    Theme by 
    Atmire NV