CS7322 Week 1 Assignment
Using Image Motion to Recognize Expressions:

Expressions are described in the paper as being a grouping of events. The `events' are described at a high level as "the lips moving up" or "the eyebrows rotating and lowering." Expressions aren't given by a static image, but by a series of images showing features transitioning to a state and then, possibly, back to the original state. At a slightly lower level, "the lips moving up" can be described by the direction certain "feature points," as the paper puts it, are moving in. These points, for example, can be the corners of the mouth, the bridge of the nose or the corners of the eyes. The paper doesn't go into too much detail as to how these feature points are extracted, but that is just a trivial engineering task, right?