This option will cause CERT to attempt to start processing live video from an attached camera as soon as it has finished launching and loading all data files. If any errors are encountered CERT will quit. This option can be used to have CERT run unattended.
This option will show the internal gray scale image, rather than the actual source image. This can be used to trouble shoot some types of problems, but can slow CERT down considerably and isn't recommended for general use.
This option creates a separate thread for grabbing frames from a live video source (e.g. webcam). On computers with more than one processor or core this can increase CERT's overall speed. Even on single core/processor machines it doesn't seem to cause any slowdown.
These two number entry fields allow you to limit the size of faces that CERT will search for. Increasing the minimum value from zero will cause CERT to not even attempt to find very small faces. Decreasing the maximum value will cause CERT to not search for the largest images. The default values of zero and twenty attempt to find faces at all scales. What a scale value means will vary depending on the size of the image, so optimum value will need to be determined through trial and error.
This popup menu lets the user choose which face alignment method will be used. In general, this should be left at the default value of 'Unconstrained Least Squares'.
This popup menu lets the user choose which face box or boxes will be displayed in the image view. The FaceFinder option will show only the original, unrotated box. The Modified option shows only the box that was used to rotate and crop the face (not necessarily a square or rectangle). The All Face Types will show both.