To provide more contextual information, we could add mean, minimum,
maximum and other aggregated values of pixels surrounding the target one. Thus
far, we computed a mean of directly adjacent pixels, but this feature could be ex-
panded to include a wider area
To provide more contextual information, we could add mean, minimum, maximum and other aggregated values of pixels surrounding the target one. Thus far, we computed a mean of directly adjacent pixels, but this feature could be ex- panded to include a wider area