Neighbourhood Grey Tone Difference Matrix

For each ROI, information about spatial changes in intensity can be obtained by looking at the difference between each pixel and the grey tones of its surrounding neighbours. A one-dimensional (1D) matrix can be estimated for each ROI, in which the ith entry corresponds to the summation of the differences between the grey level of all pixels with grey level i, and the average grey level of their surrounding neighbours. The size of the neighbourhood depends on the selected distance, i.e. a distance equal to 1 results in a 3x3 neighbourhood. Five features are derived from this 1D matrix, namely coarseness, busyness, contrast, complexity and texture strength.

+1 0

Post a comment