kus2nd's mumbling

my scrapbook before interim poster presentation ;p

Notes To reduce the effect of a noisy image on the operator output, a smoothing operator (a Gaussian blur, for example) is often applied prior to the Laplacian operator. Blurring is a process of averaging a point with its neighbors. Gaussian G_sigma(x) = (1/ sqrt(2 PI) sigma) e^(-x^2 / 2 sigma^2) Here sigma represents the standard deviation. Increasing sigma makes the shape of the Gaussian shorter and wider without changing the area under the curve (the area under G(x) is always 1). When convolved with another function, larger values of sigma provide a stronger blur. References: (accessed date:March 2, 2006) Image Processing http://www.reed.edu/~nobles/thesis/node3.html Gradient Based Method (edge detection http://homepages.inf.ed.ac.uk/rbf/CVonline/LOCAL_COPIES/MARSHALL/node28.html

posted by kus @ 6:04 PM