Collection
Data Collection using the Ultrasound with the Kinect
This manual will show how to set up and collect data from a subject. The end result will be a folder for the individual subject that holds the ultrasound video, Kinect data and coordinates, stimulus list, as well as individual frames of the tongue images.
Setting Up the Hardware Before the Subject Arrives: NEED PICTURE
Make sure the ultrasound is connected to the lockbox
The svideo cable connects to the back of the ultrasound
It is the cable that connects the ultrasound and the lockbox
The lockbox has a label for this svideo cable
Important: if it is not connected, make sure the ultrasound is off before connecting, otherwise it may malfunction
Microphone amplifier should be connected to computer
It connects in the spot with the pink ring around it
Make sure the lockbox is on
There should be a blue light on the lockbox that shows that it is on
If it does need to be turned on, press the power button on the lockbox
When you do this a box will pop up on the monitor which you should exit out of
Setting up the Ultrasound Software with the Kinect
Turn on the ultrasound
There is a power button in the top left corner of the ultrasound
INCLUDE PICTURE
Test the video feed of the ultrasound
There should be a black screen on the Titan with white grain that shows the ultrasound is on and functioning
Make sure the Kinect is connected to the computer
The USB cord connects to the front of the computer
On the main monitor, locate the ultrasound icon and double click on it
Two new windows should open, a Kinect window and an ultrasound window
Move the Kinect window to the right of the ultrasound window
On the monitor, locate the stimulus display icon
Click stimuli.py and a new window should open
Close the explorer window
The stimuli will be a group of words in any language that the subjects will be prompted with in a random order
To determine how many times these words are repeated, you will be prompted with a screen asking to specify the number of repetitions
These repetitions will repeat the same set of words in a different order each time
Click enter
CREATING A STIM LIST-NEED DIRECTIONS
A new window will open that says click to begin
Drag the click to begin window onto the second monitor and resize to fit screen
Minimize this window so you can still see the other windows
In the ultrasound window, press preview to ensure the image feed from the ultrasound is correct
Setting up Subject
Sit subject comfortably in chair
Mount microphone over subject’s ear
Apply a small strip of ultrasound gel to the top of the probe and place directly under subject’s chin to get an image of the tongue and palate
Adjust fit of probe to subject
Adjust the depth on the ultrasound to make sure the tongue is displaying correctly
NEED TIPS FOR THIS
Note: if the video appears extremely zoomed in the best solution is to turn the ultrasound on and off again
Naming the folder
In the ultrasound window, there is a space entitled recording ID
In the blank provided, include the study subject number, year, month, day, hour, and minute of the experiment in that order
Use leading zeros in the case of single digit months
Note: try not to use window’s special characters
This creates a separate file that includes the subject number and demographic data that will contain all of the data collected
Check the Kinect display
Make sure there are no obstructions in the way of the Kinect
Make sure the image feed recognizes the head of the subject
There should be a grid around the face of the subject as shown below
Collecting Data
Press start on the ultrasound window to begin recording
The record button should turn red
Note: once the recording is done it will turn grey
Click on click to begin in the stimuli window
Instruct the subject to click through the stimuli as the subject reads the words out loud
Once the subject has read through all of the stimuli, the stimulus display window should close
Hit stop on the ultrasound window to end the recording
Note: do not close this window it is very important that it stays open for post processing
The Kinect window should close when you hit stop
Post Processing and Saving Data
Post processing must be done to ensure that all of the proper data is recorded and gathered in the subject’s folder
After closing the Kinect, an icon entitled coords should appear on the desktop
Drag the coords icon into the folder you created previously
In the top right of the ultrasound window click on the post processing tab
Click Run Post Processing
Note: this may take a while, so be patient
A window will pop up that shows it is complete
Close out of this window by clicking ok
Open Stimulus Display icon again
Drag the newly created stimulusresponse.csv file into the subject’s folder that was created when you filled in the Subject ID
This file contains the stimuli, subject information, repition number, and timestamp of when the subject clicked to start each stimuli
Open Kinect Matching icon on desktop by double-clicking
Select the appropriate files for the post processing
The Kinect Matching program opens with a window that allows you to select the folders that contain the frames, video, and Kinect video and then combines them into one document labeled output
Note: Double check with each file you open that they are coming from the correct folder
Can it default to the correct folder?
Frames: locate in the subject’s folder
Video: locate the vidTimes section of the subject’s folder
Kinect: locate the coords section of the subject’s folder
Output: in the output file box, the output file will be automatically named output and placed in the same folder.
If you want it to have another name, you can fill that name into this box
Once you have located the four different parts of the folder, hit process file
A window should pop up that says file ready
Close the Kinect Matching window
Open the folder you created for the subject
A new file, entitled output should now be in the folder as well as the vidTimes, coords, and frames folders
Play the vidTimes video in order to make sure that it recorded properly
Do this by clicking the on the vidTimes file within the subject’s folder
Open the output file to make sure that the post-processing worked correctly
This should have timestamps, frame numbers, and coordinates of the locations of the subject’s head throughout the experiment as shown below