Machine learning to extract muscle fascicle length changes from dynamic ultrasound images in real-time
Data files
Jun 02, 2021 version files 8.77 GB
-
README.txt
1.01 KB
-
s1-constrained.mat
294.91 MB
-
s1-free.mat
290.25 MB
-
s1-raises.mat
289.13 MB
-
s1-walking.mat
291.45 MB
-
s1-zmix.mat
288.38 MB
-
s2-constrained.mat
296.64 MB
-
s2-free.mat
298.15 MB
-
s2-raises.mat
291.70 MB
-
s2-walking.mat
296.15 MB
-
s2-zmix.mat
291.56 MB
-
s3-constrained.mat
298.87 MB
-
s3-free.mat
294.90 MB
-
s3-raises.mat
284.59 MB
-
s3-walking.mat
288.25 MB
-
s3-zmix.mat
287.17 MB
-
s4-constrained.mat
291.89 MB
-
s4-free.mat
291.52 MB
-
s4-raises.mat
288.52 MB
-
s4-walking.mat
286.73 MB
-
s4-zmix.mat
287.35 MB
-
s5-constrained.mat
292.20 MB
-
s5-free.mat
292.41 MB
-
s5-raises.mat
291.69 MB
-
s5-walking.mat
290.27 MB
-
s5-zmix.mat
288.54 MB
-
s6-constrained.mat
301.18 MB
-
s6-free.mat
298.76 MB
-
s6-raises.mat
296.02 MB
-
s6-walking.mat
293.86 MB
-
s6-zmix.mat
295.15 MB
Abstract
B-mode ultrasound has become one-off, if not the main way of measuring muscle fascicle fiber lengths non-invasively. Yet, the gold standard for tracking these is still time-intensive hand-tracking, and even with semi-automated approaches, the process takes time and has to be done post hoc. Hence, towards greatly improving current processing capabilities by tracking these muscle fasicle lengths in real-time, we trained and optimized machine learning models with collected B-mode ultrasound data.
We focused on soleus muscle ultrasound data given the relationships existing between soleus and whole body energetics while walking and our intention to use these measurements in the loop. To ensure these data were representative of different muscle fiber loading and displacement levels, we collected B-mode ultrasound data from the soleus muscle of six participants performing five defined ankle motion tasks: (a) seated, constrained ankle plantarflexion, (b) seated, free ankle dorsi/plantarflexion, (c) weight-bearing, calf raises (d) walking, and then a (e) mix. We collected 24-seconds of 60 frames-pre-second data for each task using Telemed, ArtUs EXT-1H, LV8-5N60-A2 ultrasound probe wrapped to the right calf with 3M Vetrap Bandaging Tape. The probe was aligned so that both aponeuroses were as close to horizontal as possible in the live ultrasound video feed.
Usage notes
*** Machine learning to extract muscle fascicle
*** length changes from dynamic ultrasound images
*** in real-time
# All provided data is ultrasound data in .mat format
# The .mat files are the result after processing .tvd files
from EchoWaveII with the Matlab code provided by UltraTrack
# The .mat files can be read by UltraTrack where you'll be
able to both see the ultrasound images and do semi-automated
muscle fascicle tracking
# All data focused on soleus muscle
# Data includes 5 Tasks, 6 Subjects, and around ~24 seconds
per task
# Tasks are a) constrained ankle contractions, b) free ankle
dorsiflexion and plantarflexion, c) calf raises, d) walking,
and e) mix
# File names will include subject numbers (s1, s2, s3, s4, s5,
, s6) and tasks (constrained, free, raises, walking, zmix)
For example: "s4_walking.mat"
# All data we used is included (all 5 tasks for each of the
6 subjects)