Example showing gct_Tahltan plugin applied to Xplane (simultaneous midsagittal and coronal) image of /ae/. Red dots are interactively adjustable spline anchor points controlling tongue contour fit at right, and defining medial groove depth at right.
GetContours is a Matlab-based platform that provides straightforward click-and-drag positioning of reference points controlling a cubic spline fit to a displayed ultrasound image of the midsagittal tongue surface. This fit – the “contour” – provides a discretized representation of the tongue suitable for subsequent analysis. Contour fitting may be either interactive, driven by image forces, or automated through sequences. Features of the program include:
-
support for loading from and navigating within DICOM sequence and movie formats
-
tools for image filtering, averaging, and contrast enhancement
-
Praat TextGrids labeled on associated audio can be imported to identify and annotate articulatory events of interest, allowing rapid selection of key frames within image sequences
-
audio from the source movie is displayed if available and may be played to provide context for the displayed frame; clicking on the audio plot may also be used to set the active frame
-
support for image pre-processing on loading (e.g. cropping, resizing, or custom enhancement)
-
optional ‘plug-in’ procedures support automated tracking through sequences (e.g. gct_Lines, gct_snake, gct_SLURP)
-
incorporates Cathy Laporte's implementation of the EdgeTrak (Li et al. 2005) ‘snake’ active contour model for fitting the tongue surface using image forces
-
provides tools for automatically identifying mm/pixel scaling factor and coordinate origin (center of probe arc circle), supporting the export of mm-scaled coordinates relative to that origin
-
contours may be exported to tab-delimited text files, and converted to wide or EdgeTrak formats using ReshapeContours
-
support for estimating Fourier fit coefficients from contour shape (Liljencrants 1971)
-
separate DotsPlace, DotsTrack procedures can be used to track image features from co-collected video to generate e.g. head-centric correction signals for extracted contours
-
the ComputeCurvature procedure provides curvature-based measures of the number of inflection points (NINFL) and the Modified Curvature Index (MCI) for the current frame
Example showing result of automated tracking using gct_SLURP, showing best-fitting contour at right with superimposed interactively adjustable spline anchor points, and fit diagnostics at left
Example illustrating Fourier coefficient shape fitting
See help GetContours within Matlab for usage and examples.
UltraFest2015.pdf (a poster presented at the Hong Kong Ultrafest VII) provides an overview of (an early version of) GetContours.
Please contact mark tiede with questions, problems, or feature requests. Many thanks to Cathy Laporte for the snake and SLURP plugin implementations.
Release Notes:
v0.8 09DEC15 (UltraFest HK version)
v0.9 25AUG16 bug fixes, Snake implementation, Dot tracking
v1.0 22OCT16 inherit anchors into previously visited empty frame
v1.1 11NOV16 fix explicit keyframes issue
v1.2 02FEB17 fix point addition order, DotsTrack update
v1.3 14JUN17 support skipping bad regions in DotsTrack
v2.0 05JUN18 support sequence tracking, tracking plugins
v2.1 15JUL18 fix TextGrid issues
v2.2 01OCT18 fix initialization overwrite bug
v2.3 12NOV19 mods for internal improvements; use SLURP snake implementation
v2.4 03MAR20 bug fixes, scroller, DICOM, SLURP support
v2.5 12AUG20 bug fixes, added gct_snake (UltraFest 2020 version)
v2.6 19AUG20 support for Fourier coefficient shape fitting
v2.7 26AUG20 support preseeded anchors
v2.8 15SEP20 fix DICOM close
v3.0 08SEP20 support audio panel
v3.1 15SEP20 support draw mode, multiple panels
v3.2 03OCT20 support info, frame differencing, anchor deletion issue
v3.3 22OCT20 bug fixes
v3.4 23OCT20 DICOM file extension issue, cfg bug, mark editable GC window
v3.5 28JAN21 fix gct_SLURP mask initialization; add measurement tool, slowdown with modifier keys on cycle
Copyright (C) 2015-2021 mark tiede
Publications which make use of gct_SLURP should cite Laporte, C. & Ménard, L. (2018).
Multi-hypothesis tracking of the tongue surface in ultrasound video recordings of normal
and impaired speech. Medical Image Analysis, 44, 98-114.
This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, version 3 or any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.