From: Antonio Garrido Carrillo <agarrido@ug...>  20070201 18:50:14

Another quick question.... I have been taking a look at vil_suppress_non_max_edges.txx, where a method for non maximal suppession is implemented. It's very simple, but I have a question because I have been working with methods like this many years ago. In order to decide if the point x is a maximal edge, we compare magnitude at x (mag) with scalar product of x and gradient at x+(grad_i/mag,grad_j/mag). Probably, the result is almost the same, but another possibility is compare magnitud at x with magnitude at x+(grad_i/mag,grad_j/mag). In this case, only a call to vil_bilin_interp_unsafe is needed. Well, any opinion about this? thank you 