A Qualitative Analysis of a Novel Face-Tracking Application and a Proposed Research Design to Study Glossectomy Speech using this Method

  • Author / Creator
    Khaledan, Roujan
  • Background: Following glossectomy, patients often utilize compensatory articulatory behaviors (CABs) to compensate for their acquired speech impairments. Incidental findings have indicated the presence of lower lip inversion CABs, but no study has investigated these inversions directly. A prototype face-tracking application was developed in-house at the University of Alberta that had the potential to be used for the proposed CAB study. However, due to its novelty for use in research, a qualitative analysis was needed to demonstrate how whether the prototype could capture the data necessary for the proposed CAB study.
    Objective: The objectives of this thesis were to: 1) design a cross-sectional matched control study to determine the extent of lip inversion in glossectomy speakers when producing alveolar consonants; 2) determine the feasibility of the prototype face-tracking application for capturing speech kinematics; 3) investigate potential sources of error under different recording conditions.
    Method: Objective one (designing a study) was addressed by designing a cross-sectional matched control trial to characterize lower lip CAB for alveolar sounds in glossectomy speech using the prototype application. This study was designed to be run with 20 participants in each of the experimental (glossectomy) and control (healthy) groups. Experimental words had voiced or voiceless alveolar sounds either word initially or finally. Objective two (determining the feasibility of the prototype for research) was evaluated by measuring various moveable and static portions of a single healthy participant’s face and plotting movement trajectories during speech and non-speech tasks. Objective three (investigating potential sources of error) was investigated by assessing the standard deviations (SD) of select points on the face at rest under different recording conditions; the error during a speech task as also investigated by assessing the SD of a static point.
    Results: The proposed research protocol was designed to compare lower lip movement of glossectomy speakers during alveolar sound production to healthy speakers by using a mixed-effect linear model (α=0.05, two-tailed). Then, the qualitative analysis showed that the prototype face-tracking application captured movement of articulators in speech and non-speech tasks. Also, the SD of measured points across all recording conditions were ‘highly’ accurate.
    Conclusion: The work from this thesis provided a method for future researchers to study CAB in glossectomy speech. It also demonstrated the feasibility of the prototype for use in the proposed CAB and as well as other speech kinematic studies.

  • Subjects / Keywords
  • Graduation date
    Fall 2021
  • Type of Item
  • Degree
    Master of Science
  • DOI
  • License
    This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.