I wonder if it would be possible to filter pixels (similarly to this project) from the screenshot (EDIT: yes, the border around the two words) to make them brighter?
Any ideas?
p.s.
Below is the result to compare with original above:
I'd try converting the pixels of the costume from RGBA into HSVA representation and then doing your filtering based on those criteria
1 Like
Well, first try to make your criteria slightly less fuzzy. I take it we're talking about the cyan pixels to the right of the black pixels? Like, the cyan vertical to the right of the letter l?
I imported your picture to Photoshop to look at the color values. (You can do it in Snap! too, since we have the "color at" block.) The leftmost l (that's a lower case L) is
R 56, G 111, B 146; H 203, S 62, V 57.
The second one is
R 84, G 141, B 171; H 201, S 51, V 67.
The third and last one is
R 62, G 127, B 170; H 204, S 64, V 67.
So, I agree with @cymplecy that the obvious criteria are H∈[195,210] and V∈[50,70]. This is too bad, because what you're given is an array of RGBA values, not HSVA. But you can use the Colors library blocks to convert: First
with the pixel (a list of four values) in the open slot. Then use
or individual values 
, 
to do the testing.
1 Like
You want to process blue border? Or the area inside?
And You do not try to recover dark pixels under border?
1 Like
@cymplecy Thank you for the general guideline about converting RGBA to HSVA.
@dardoro Yes, only the 'border' (around the two words); sorry for not giving a precise description to start with.
@bh Thank you for your suggestion to use the blocks from Colors library.
Below is, more or less, what I was aiming for.
I have to admit I've arrived at the criteria producing the results, again, more or less using trial-and-error method ;~P
