STATISTICAL FOUNDATION FOR HYPOTHESIS-TESTING OF IMAGE DATA

A statistical foundation is given to the problem of hypothesizing and testing geometric properties of image data heuristically derived by Ka natani (CVGIP: Image Understanding 54 (1991), 333-348). Points and lin es in the image are represented by ''N-vectors'' and their reliability is evaluated by their ''covariance matrices''. Under a Gaussian appro ximation of the distribution, the test takes the form of a chi(2) test . Test criteria are explicitly stated for model matching and testing e dge groupings, vanishing points, focuses of expansion, and vanishing l ines. (C) 1994 Academic Press, Inc.