Effects of border variations due to spatial quantisation on binary-image template matching

Quantised binary images are formed by projecting the binary intensity silhouette or line edge of a shape onto a spatial quantisation grid. The quantised shape depends on the orientation and translation of the object´s image. Different complexities in the borders, due to the quantisation effects, are part of the error arising in a binary template-match process. The letter suggests a model to approximate the variation statistics as a result of uncertainty in translation only. To isolate the quantisation effects, noise-free images are considered.