Generate an image: An N × N image consists of N2
pixels. We can choose the grey-level brightness for each pixel from almost any distribution
provided that the chosen value is real and non-negative.
What is the usual basis for requiring non-negativity?
In the following experiment we choose the brightness from one of seven possible distributions.
We then generate an N × N image.
Use the slider to generate an image with N2 pixels using the selected “mystery”
distribution for the grey levels. Radio buttons allow you to 1) choose the underlying distribution
and 2) display either the histogram of brightnesses, the estimated probability density/mass
function, or the cumulative distribution function.
What is the conventional name for each of these distributions?
How many pixels did you need to identify the distribution with a reasonable amount
of confidence?