How can I find a good image based on a histogram

Sorry, I'm using the word "good" ... it's relative. But in my project I want to find photos that are not under or overexposed and have great dynamics with a lot of colors.

okay:

  • for the first tasks, I can take a V-histogram in the HSV color space with a rough number of bins. Therefore, I can detect under / overexposed images
  • can I use an H / V-Diagram to count colors, perhaps in a cumulative histogram?
  • For dynamic (the difference between the speed of colors) I can select the H-Diagram and calculate avg from the bins.
  • In addition, I want to detect global peaks. Where the color is blurred. Of course, I can get the maximum of the V-histogram and see if the value / brightness is 255.

These points represent my ideas for solving problems. But maybe there are other, effective solutions or ideas from any documents.

In my project, I have to research a large number of images, and I think it takes a long time to create all these histograms.

Hi

+4
source share
1 answer

Try the Quantification Colors - I'm sure it will help you. Have a look at this discussion fooobar.com/questions/1422414 / ....

+1
source

All Articles