Recognition of objects with skew distortion based on synergetics

A synergetic approach to recognition of objects with skew distortion is presented. This kind of deformation is usually found in Vehicle License Recognition Systems due to the specific locations of video capture devices. The shape deformation is integrated in a dynamic system derived from the basic synergetic equation, and the transformation is swapped from the prototype patterns to the test pattern so that the structure complexity is reduced. A low-pass filter acting on the potential function is introduced in order to eliminate sharp fluctuations that may lead to pseudo-traps while the overall shape of the potential curve is maintained. The method above can also be applied to invariance problems including translation invariance, rotation invariance, scaling invariance and so on. (C) 1999 Elsevier Science B.V. All rights reserved.