Abstract
It is often required in the analyses of the eye movement to divide the observed data into the saccadic and smooth components. Several computer algorithms have been proposed so far, which used the first- or second-order derivative of the movement with respect to the time for descriminating the two components. This paper describes a new method of using the third-order derivative. It separates the two components more clearly and is more insensitive to noises than the previous methods.
