Reputation: 38063
I wish to interpret potentially monochrome bitmaps which appear to have been antialiased or compressed through JPEGS. Here is an example:
Although it appears monochrome magnification shows:
Where have the colours come from? If we assume the original was monochrome (black + white) with sharp edges can we use this information to reconstruct the original (with less loss than simple image analysis)? If so is there a F/OSS library (Java preferred) to do it?
NOTE: I have altered the title because of @duskwuff's answer
Also the non-text seems to be grayscale (as implied by @duskwuff). Is this antialiasing? and can it be reconstructed?
Upvotes: 0
Views: 137
Reputation: 3135
Where do they come from? The original was most likely not really monochrome. It is common to use varying shades of grey to help reproduce a straight line. Zoomed in it may not look the best, but at the intended display scale this makes the pseudo binary image look better. Did you look at the original to confirm it is really monochrome?
Remember you are looking at a sampled version of the line. If they sample using linear interpolation (will look like what you have) it looks better then if they had done nearest neighbor (will be strictly monochrome).
If you wish to convert to monochrome you need to threshold the image to make it binary. Simply loop through each pixel and if the value is <128 make it zero and if it is >=128 make it 255.
Upvotes: 1