2 days ago
Put the paintbrush down – AI can restore artworks quicker and better
Artificial intelligence (AI) could spell the end of art restoration by humans after MIT showed that damaged paintings can be repaired in just a few hours.
Typically, conservators spend months or years researching and matching paints, colours and techniques to ensure the finished product is as close to the original as possible.
But Alex Kachkine, an engineering graduate at MIT, has shown it is possible to use AI to fill in the damaged areas digitally, then print the restored layers onto a thin film to attach on top of the painting.
It means the painting appears restored even though the original is still intact beneath.
The method was applied to a highly damaged 15th-century oil painting, and AI immediately identified 5,612 separate regions in need of repair, and filled in these regions using 57,314 different colours. The entire process, from start to finish, took just three and a half hours.
'There is a lot of damaged art in storage that might never be seen,' said Mr Kachkine. 'Hopefully with this new method, there's a chance we'll see more art, which I would be delighted by.'
In recent years, digital restoration tools have allowed conservators to create virtual representations of restored works.
AI algorithms can quickly sift through huge amounts of data about artists and time periods to generate a digitally restored version of a particular painting, in a way that closely resembles the correct style.
However, digital restorations are usually displayed virtually or printed as stand-alone works and until now there has never been a way to translate the digital restorations on to the original work.
The new technique involves first scanning the painting and then using software to create maps of the areas that require repair, before matching the correct colours.
The maps are then inkjet printed onto two sheets of a film. One film, which sits underneath, is printed in white so that the colours overlaid on the second sheet are not muddied by the colours of the original painting.
The layers are carefully aligned and overlaid by hand onto the original painting and adhered with a thin spray of varnish.
The printed films are made from materials that can be easily dissolved with conservation-grade solutions, in case conservators need to reveal the original, damaged work. The digital file of the mask can also be saved as a detailed record of what was restored.
For the painting that Mr Kachkine used, the method was able to fill in thousands of losses in just a few hours.
'A few years ago, I was restoring this baroque Italian painting with probably the same order magnitude of losses, and it took me nine months of part-time work,' he said. 'The more losses there are, the better this method is.'
The new technique may even help restore botched restorations such as Elias Garcia Martinez's Ecce Homo, which was restored so badly it was dubbed Monkey Christ, because of the Simian features of Jesus.
Mr Kachkine acknowledges that there were ethical issues to consider, in terms of whether a restored version was an appropriate representation of an artist's original style and intent.
He said the process should be carried out in consultation with conservators with knowledge of a painting's history and origins to check the AI programme was not going off-piste.
'Restoring a painting is fun, and it's great to sit down and infill things and have a nice evening,' he added 'But that's a very slow process.'
The research was published in Nature.
Broaden your horizons with award-winning British journalism. Try The Telegraph free for 1 month with unlimited access to our award-winning website, exclusive app, money-saving offers and more.
