At 01:34 PM 10/27/02 +0100, Eugen Leitl wrote: > (Try [to] rewire parts of CNS in control of your motorics).
Actually, injuring yourself is a good way to do this. You'll screw up the timing differences (albeit not the geometry) royally. We call this a limp. If you're interested in gait perception, look up the psychologists Michotte, Johanssen, Cutting; "point-light-displays". Its been known you can discriminate sex on the basis of a handful of movie frames and only a dozen visible points. (Based on human sexual dimorphism and biomechanical corollaries in gait, e.g., women have wider hips, and men swing their shoulders more.) And one can tell if a point-light figure is carrying something heavy or light. One can, by varying parameters in synthetic displays, generate recognizable instances of the same gait (walking Elmer Fudd, Mae West). It is not rocket to science to see that by visually recovering such parameters from real videos one could recover a set of parameters that are reasonably invariant and are a signature of an individual. Funky lissajous traces of a button on your butt. Look into biomechanics and the encodings of gait as duty-cycle and phase. Look up the photographer Muybridge. Read the Pixar folks' papers on animating motion. Papers on structure from motion, recovering 'events' in moving images. Papers on diagnosing motor impairments in children.
