The authors give a possible solution to a problem of how to combine effectively the different scales of analysis at the early visual information processing level. The main result is a limit theorem which confirms that in 1-D the convolution difference uniformly and asymptotically approximates the box train function as the size of the Gaussian filter approaches zero. The box train function represents the edge point information in the 1-D image. The modified limit theorem for detecting edge directions is presented.
|Original language||English (US)|
|Title of host publication||Unknown Host Publication Title|
|Number of pages||6|
|State||Published - 1986|
ASJC Scopus subject areas