python - Simple Image Metrics with PIL -

I want to process uploaded photos with PIL and determine some "soft" image matrix like:

• Is the image opposite or monotonous?
• Color or monochrome?
• Bright or dark?
• Image is hot or cold (about light temperature)?
• Is there an impressive color?

Metrics should be measured in rating-style, e.g. Color ++++ for a very colorful photo, rather colorful + for a monochrome image.

I have already seen PIL's IMSTEST module, which calculates some interesting values ​​for my metrics, e.g. Histogram etc. RMS However, this module is poorly documented, so I'm looking at a more solid algorithm to determine these metrics.

I do not think there are methods that give you a metric you want, but the way Is near, like RMS, shows you a long way there. To do things with color, you can split the image into one layer per layer, and you can get RMS on each layer, which tells you a few things that you want to know, To convert so that you can only retain color information, etc.