The application of a mathematical operation to each neighbourhood in an image is called convolution. The operation is defined by a “mask” specifying for each neighbourhood, how many points it contains and how the corresponding image point affects the computations. Each location in the operator mask contains a weighting value, these are multiplied by the value of the corresponding image location and the results summed to give the convolution value for that neighbourhood. Doing this for all neighbourhoods produces a new array of values. Mathematically, the convolution integral is the integrated cross product of a weighting function with an image. See local grey-level operations <127>.
KeywordsData Item Image Point Modulation Transfer Function Mathematical Operation Normalise Description
- [Frisby 79]Frisby, J. P. Seeing. Oxford University Press, 1979.Google Scholar