Oct.15-Oct.??, 2003

gives a random order of frames (samples?)

        (1) achieve better encoding

        (2) achieve a particular false alarm rate.

One way to evaluate this is to pick up a 'leave-one-out' frame.  Test pixels in the frame against the nearest codeword.  If accepted, calculate a RMS distance over the pixels on that frame.  Otherwise, count 'not-accepted' to get a false alarm rate.  Here, I got a question.  Does the RMS distance represent 'color distance'?  What about 'brightness distance'?