i have an image and would like to perform a canny edge detection. one of the first steps is the convolution of the image with a smoothed derivative filter, i.e. a gaussian. the problem is, that i don`t know, how to convolve my collection of discrete pixels (the image) with a 2D gaussian. how do i get an integer-valued convolution kernel that approximates a Gaussian with a variable sigma?

Hey there,
i have a table of points that describe a rectangular function and i am trying to calculate the middle of the maximum or minimum (depending on the function) of this datacloud. my first attempt was to spline the data and then use the minimize command, however, i am not sure if the minimize command gives me the right result because it depends on my interval length. the second attempt was to calculate the derivative of the spline. but how do i get then the exact middlepoint?...

I have an BW image, which i can read in with ImageTools. now i simply want a histogram of it. the plothistogram function works, but i am not sure that it does right, because the numbers in the plot seem far to big. so i am trying to use the histogram function but i am only able to plot a histogram for a single vector of my image. is there an easy way to use the histogram function for the entire array? as the next step, i would like to compute the variance of the histogram....