1. [25] Segmentation techniques based on spatiotemporal locality are doomed to fail in most complex tasks of speech and vision understanding. However, they can be useful in restricted application tasks — a good reason for which they still deserve attention. The segmentation of speech bursts can be based on checking the presence of “silence” in a given utterance, that is, checking when v2(t)  0. For video signals, a simple segmentation, which aggregates regions with uniform gray level, can be driven by checking (∇xv(t, x))2  0. For a speech signal, one might think of the dual condition ˙v 2(t)  0. Why such a duality is not appropriate as a segmentation scheme?