Is it possible to represent a characteristic of an image as a single value?

Hello there.

Said characteristic would be luminance. The idea would be to have specific thresholds classify different kinds of images as either dim, normal, or bright. My idea was to calculate the intensity of an image, and compare it to the thresholds found experimentally, but I failed to realize then that the intensity isn't represented as a single value, rather as a matrix of pixel values.

Is it possible? Thank you in advance for your replies.

Answer 1

[Adam]

If you need a single value for the whole image then mean, median and other similar statistics across all your pixel values will give you that.

Alternatively you could apply the thresholds individually to each pixel and then determine your classification based on what proportion of the pixels fall in the range of a given class.

[Cosmic Clarity]

I thought of that second idea, actually, but that would possibly mean a long processing time. However, that first idea may work. I thought of using just the mean initially, but thought it wouldn't be accurate enough using just that.

Thank you for the reply. Both ideas will help a lot. :)