CERT is an application for automated FACS. The Facial Action Coding System (FACS) is a way to categorize facial behaviors based on the muscles that produce them.

The basic processing steps for each frame of video, or individual image is:

1. The image is given to an automated face and feature detector
2. The face is cropped and rotated based on feature positions
3. Then passed into the action unit coding system
4. Depending on which set of weight files are used a number of different AU's will be calculated. For example:
   • AU 1: Inner brow raise
   • AU 2: Outer brow raise
   • AU 4: Brow corrugator
   • AU 5: Upper lid raise
   • AU 10: Nasolabial furrow, "disgust"
   • AU 12: Lip corner pull, "smile"
   • AU 14: Dimpler
   • AU 20: Lip stretch, (pull down corners of the mouth) "Eeek!"
5. Each output value gives a relative indication of how much a particular AU is present.

For more details about the AU processing, see this page of the main MPLab web site.

After launching CERT you should see the main window with the following parts:

• The MPLab logo at the top with buttons and a text field below it.
• The image view is below the buttons, on the left.
• Depending on which plugins are enabled, there may be a progress indicator showing the status of loading data files.
• Also depending on which plugins are enabled, there may be some data views (graphs) to the right of the image view.
• Below the image view is a box with more controls.

In addition to the main window there is a Preferences window with general preferences along with plugin related items.

There is also a Face/AU Data window that is used for viewing or saving CERT output.

A typical session of using CERT might go like this:

1. Launch CERT.
2. Using the Plugins menu or the Plugins pane of the Preferences window, choose which plugins should be enabled.
3. Go to the Face/AU Data window and click on the File... button and choose a location for saving the data. Click the Start Recording button.
4. Wait until all data files are loaded (the progress indicator will disappear)
5. Click on the Choose Source... button and choose an image, movie file or folder containing image or movie files. Or click the Live Video button to use an attached video or web camera.
6. CERT will load each image or movie frame and process it. All data will be saved to the chosen file.
7. Click the Stop Recording button in the Face/AU Data window.

A few features (most are covered in more detail elsewhere):

• Holding down the option key as CERT launches will prevent the loading of data files, and some other automatic behaviors.
• Clicking on a data view will cause it to grow, which can be useful when giving demos so viewers can concentrate on one AU.
• Some plugins may supply a contextual menu. Control-clicking on the plugin's data view(s) will show the menu.
• The size of video from an attached camera can be adjusted, before starting, using the Width and Height fields in the box below the image view.
• A preference determines how many decimal places are used to output data. There is no user interface to change it, but the following command could be used in a Terminal window to change the number of decimal places to ten: defaults write ~/Library/Preferences/edu.ucsd.mplab.AUCoderDemo.plist decimalPlaces '10'
• The output of Face Finder and all Feature Detector plugins is displayed overlaid on the face in the image view, rather than in separate data views. The output is also included when recording data to a file or to the Face/AU Data window

• Preferences
• Data Window
• Plugins