Perhaps this gives them more control over the contents of the image.
You can analyze the color range of the image, the prevailing color, even its contents (taking into account the smart algorithm).
You can also make pixel-by-pixel changes - change brightness, contrast, gamma, etc. (not sure why they would like to do this, perhaps for some future use).
You can also rotate the image (in browsers that support canvas, but not in CSS transforms, see for example this demo of my fabric.js ).
kangax
source share