Graphic Design Stack Exchange is a question and answer site for Graphic Design professionals, students, and enthusiasts. It only takes a minute to sign up.
Sign up to join this community
I have a very very old photo of a person that I want to color it and make it natural. First of all I used a fft action to remove bubble algorithm of the photo. Then I tried to make that photo colored but it not seem very natural and nice. What extra works can I do to have a beautiful photo.
Original photo
After using fft action
Brushing colors in a separated layer
Make that layer multiply, fill 70%. I did the same work for hair, eyes, etc.
Your efforts have already produced as good result as possible without inserting details. Actually you already have inserted some colors without knowing the real color.
Your image is blurry. Sharpening can make it better looking, but it very easily becomes more like a drawing, it doesn't add missing information and noise jumps up. An example by https://pinetools.com/sharpen-image
Dust removal techniques such as median filtering or some more clever tools can guess sharp edges and keep them, but suppress some sharp small spots, Subjectively the result again can be a little more better looking. An example by https://www167.lunapic.com/editor/
Filtering your image doesn't add the missing details. You must add them if you need them. You can try to warp an existing nearly same face to the same geometric form or take the missing detail from several different sources.
If you find a good way to blur and otherwise detoriate your augmented photo temporarily, you can see if the added details fit.
Finally the result can be extremely plausible if you are patient enough.
But that's not the original face, it's your guess.
Intelligence and forensic investigators surely have huge libraries of facial features and some clever software. They can find the most probable library details which can result the existing not so sharp photo. Combining them obviously makes what you want. It's not the same as filtering your photo, it's a statistical pattern matching job to deduce the most probable face.
I have no idea how ordinary people could do the same.
