In Colors theory, what is the difference between saturation and brightness?
There's no point rattling off another 1-line answer because it's a rather large topic.
Wikipedia has this for HSL and HSV. Next there's an brief tutorial on Color Theory. Finally more on HSV And H2SV Color Space with some transformation code in C++.
Related
Wikipedia and plethora of online resources provide detailed and abundant help with various color space conversions from/to RGB. What I need is a straight YUV->HSL/HSV conversion.
In fact what I need is just the Hue (don't care much for the Saturation or the brightness Lightness/Value). In other words I just need to calculate the "color angle" for a given YUV color.
Code in any language would suffice, though my preference is C-style syntax.
Note that by YUV I mean specifically Y′UV, a.k.a. YCbCr (if that makes any difference).
While YUV->RGB colorspace conversion is linear (same as "can be expressed as a matrix operation") the RGB->HSL is not. Thus it is not possible to combine the two into a single operation.
Thank you Kel Solaar for confirming this for me.
For reference:
YUV(YCbCr)->RGB conversion
RGB->HSL conversion
Note that mathematically the calculation for Hue is written piecewise as the "base angle" depends on which sector the color is in and the "major color" is driven by the max(R, G, B) expression.
I think they are from the different worlds of interest. Here is a google patent
https://patents.google.com/patent/CN105847775A/en
I'm implementing the Delta CMC algorithm (color difference in CIELAB space) as described here and here.
I was surprised to see a calculation for the Hue of the first color but not the second color. This would most likely make the algorithm asymmetric. Is this right?
Would it be better to average the Hue of the two colors and use that instead?
You are right, CMC algorithm is really not symmetric. Quote from here:
the formula is not symmetric, but depends on a standard and a test colour.
Switching standard and test colors changes the difference.
Assume that I've a gray image, and I want to draw e.g. text on it. Now the image has some dark and some bright regions. So if I choose for every character a separate color, in what way do I compute such a color to gain highest contrast of the text?
A pragmatic approach is to use yellow. (I don't know why, but its often used for subtitles in movies and documentaries)
Furthermore I could darken the yellow in regions of bright background, and highlight it in regions of dark background. But this may provide some layer-effects.
I know that the color space may be important. I start with an RGB gray value, but LAB, HSV, or HSL may be better suited to compute the optimal color.
EDIT:
As there was the question for a useful use-case: I really do not want to paint letters of text in different color. It is about color choosing for particular glyphs on gray textured background. (E.g. an MR image.)
The simplest answer to your question is to maximize the distance between the background and your text color.
If you convert to HSL, you can do this by maximizing the distance between L (V in HSV). And all that requires is to select white when the background lightness is less than 50% and black otherwise. Here is an example: http://jsfiddle.net/E2kU4/
if(bgLightness < 50){
color = "white";
}else{
color = "black";
}
I think that pretty much solves it, but on to a few other points:
I'm not sure what the use case is for this. A word with different colored characters might look really bad. Typically, subtitles select a single color for consistency.
Yellow does stand out against a black and white image because of its saturation. Furthermore (and I'm not sure how to put this into words exactly), yellow has a really high chroma compared to other colors with similar lightness. It is best demonstrated on the HUSL page; by the way, HUSL is a great library for creating readable colors.
Yellow easily contrasts with dark colors because it is very light. It doesn't contrast with light colors as well, but that is usually solved by adding a shadow/outline in subtitles. Another example: http://jsfiddle.net/E2kU4/1/
But you can apply the same technique (of applying a shadow or outline) to the black/white example for maximum contrast. Now, the outline has the maximal contrast against the text. The outline stands out from the background too, unless those colors are similar, in which case the contrast is already extremely high (e.g. Near black background, black outline, white text) http://jsfiddle.net/E2kU4/2/
Lastly, converting to and from HSL/RGB should be trivial. There are plenty of libraries to do it.
I've heard that if you need to do a color segmentation on your software (create a binary image from a colored image by setting pixels to 1 if they meet certain threshold rules like R<100, G>100, 10< B < 123) it is better to first convert your image to HSV. Is this really true? And why?
The big reason is that it separates color information (chroma) from intensity or lighting (luma). Because value is separated, you can construct a histogram or thresholding rules using only saturation and hue. This in theory will work regardless of lighting changes in the value channel. In practice it is just a nice improvement. Even by singling out only the hue you still have a very meaningful representation of the base color that will likely work much better than RGB. The end result is a more robust color thresholding over simpler parameters.
Hue is a continuous representation of color so that 0 and 360 are the same hue which gives you more flexibility with the buckets you use in a histogram. Geometrically you can picture the HSV color space as a cone or cylinder with H being the degree, saturation being the radius, and value being the height. See the HSV wikipedia page.
I'm looking for an idea for getting the most representative color in a grid of pixels. There is any algorithm for this? I'm not sure if the most representative is one of the colors appearing in the grid of is the average af all the pixels better?
alt text http://www.stan.mx/images/stackoverflowPixels.gif
Have a look at some color quantization algorithms. I found them to be the most effective method to generate palettes from photographs. Also, most image manipulation/processing libraries should have some fast quantization built in.
You are probably looking at "average" as percepted by human. First you need to change you colors representation in a color space that is specially designed to be
"perceptually uniform" (for calculation of color "distances") Lab* link text
Then, each color is a point in 3D color space. Now you can find the "center" of the cloud of points and this is the "most representative color".