-
Operation based on transformation of a contour to signal (signature) by simulation of a phenomenon of narrowing and expanding field of vision (currently a sample (signal) of 128 bytes, irrespectively from the level of complication of the contour
(„first glance”))
- „Resistant” to defects in contours connected e.g. with shakes of hand (recognition on the basis of drawings by hand) and breaks in the outline of an objects (discontinuous patterns), or imperfections (e.g. partly concealed object)
- Fast (possible to work on a real video image)
- NOT based on neural circuitry (no circuit training phase – a new object is added to the base by its index and a reference to a real object)
- DOES NOT use Fourier transformation
or fuzzy logic
|
|
| |
