Photo Restoration with AI, Done Carefully
Repair and colourise old family photos while keeping faces and history true
A taste of a lesson
I used an AI restore button on my grandmother's 1940s photo. It is sharp now, but she looks like someone else. What happened?
The tool did not recover her face; it rebuilt one from patterns it learned from many other faces, and some of her features were replaced. That is common with strong face restoration. Go back to your original scan. Try the restoration at a lower strength, or blend the restored layer at about 30 to 50 percent over the original so her real features show through. Compare eyes, mouth and jawline side by side. What looks most different to you: her eyes, her smile or her face shape?
Written by the teacher as an example. In your lesson the tutor answers your own questions, and like any AI it can be wrong.
What you will be able to do
- Scan old photos well and keep an untouched master
- Choose the lightest repair method for each type of damage
- Check restored faces for invented details and identity changes
- Label restored and colourised images honestly
Lesson plan
- 1 Scanning and the master copy Capture the photo well and protect the original file. Start
- 2 Light repairs first Fix dust, scratches and fading with the gentlest methods. Start
- 3 Faces: recovering or inventing? Use face restoration carefully and spot identity changes. Start
- 4 Colourisation as interpretation Understand what colourisation guesses and how to improve it. Start
- 5 Labelling, sharing and privacy Share restored photos honestly and respectfully. Start
Try asking
About this tutor
For family historians, archivists, and anyone with a box of damaged old photographs. AI can repair scratches, tears and fading and can add colour, but it can also quietly invent faces, change expressions and add details that never existed. This tutor teaches a careful process: scanning well, keeping the original untouched, choosing the lightest repair that works, checking every face against the original, treating colourisation as an interpretation, and labelling restored images honestly so family records stay accurate. Gentle and patient, for complete beginners.
Reviews
4.6
5 ratingsSample
- Rajesh V.Sample
The scanning lesson alone was worth it. I was scanning at low resolution as JPEGs. Redid the whole album properly.
- Felipe A.Sample
Explained clearly why my great uncle's uniform came out the wrong colour, and how to research the right one.
- Dorothy H.Sample
Patient and kind. I learned to blend the restored face with the original and now my father looks like my father again.
- Grace O.Sample
Simple steps, no jargon. Would like help with very damaged photos with large missing pieces.
- Nora B.Sample
Good advice on labelling colourised photos for our family tree. A bit cautious about AI but I understand why.
About the teacher
Illustration, picture books, comics and game art with AI, with the artist kept in charge
9 tutors 396 lessons taught Sample
I am an illustrator who teaches other illustrators, writers and curious beginners how image models fit into drawing and storytelling. I work mostly on picture books, comics and concept art, and I use AI for exploration, reference and pattern work while keeping the final decisions and much of the hand work my own. I teach by looking at pictures together...
See Pilar's profile and tutorsMore like this
Other tutors on the same or nearby topics.