CASL
Permanent URI for this collectionhttps://eresearch.qmu.ac.uk/handle/20.500.12289/22
Browse
2 results
Search Results
Item Improving ultrasound post estimation accuracy by training on co-registered EMA data(University of Aizu, 2024-06-28) Balch-Tomes, Jonathan; Wrench, Alan A.; Scobbie, James M.; Macmartin, C.; Turk, A.This study aims to assess how accurately DeepLabCut [1], when applied to ultrasound tongue images, can estimate Electromagnetic Articulography (EMA) sensor positions. EMA provides objective measures of anterior tongue, jaw, and lip kinematics. DeepLabCut pose estimation is a powerful method of extracting keypoint positions from midsagittal ultrasound images of the tongue. It has an advantage over EMA in that it can be applied to the whole of the tongue from tip to root as well as the jaw and the hyoid. After correction for probe translation standard error in the estimation of keypoint positions compared to the corresponding EMA sensor positions was 1.2-1.5mm along the tongue contour and 0.5-0.9mm perpendicular to the tongue contour.Item Similar and different tongue surface contours: intra-speaker controls in ultrasound analysis(2023-08) Scobbie, James M.Ultrasound studies of speech production analyse differences in dependent variables reflecting the tongue surface’s location and shape. Inferential statistics distinguish theoretically-relevant from random effects, somewhat independently of the descriptive size of significant effects. Experimental designs induce measurable dependent changes by manipulating independent variables such as prosody, phonemic target, etc. This paper presents descriptive statistics quantifying holistically all 15 pairwise differences between six monophthongal long vowel phonemes of one variety of English, comparing these to experimental noise differences attributable to the use of two identical blocks of data collection in sequence. Eight speakers were recorded, using two different ultrasound systems, and analysed in AAA using both edge-tracking and DeepLabCut pose estimation. The smallest phonemic contrast (~2mm) was greater than the experimental noise (~1mm), and was well evidenced by AAA’s t-test of radial difference.