Massachusetts Institute of Technology recently announcedAn artificial intelligence application technology, it will be possible to perform "reversible" restoration work on historical paintings and achieve more accurate restoration results.
This technology, proposed by mechanical engineering graduate student Alex Kachkine, uses an artificial intelligence algorithm to analyze the content of the proposed restoration and build an artificial intelligence model that can restore the painting's original appearance. It also uses a registration algorithm based on the principles of human color perception to generate a "damage map." While ensuring pixel-level accuracy for each layer, it identifies the locations on the painting where color is peeling, cracking, or fading, thereby generating information related to the colors that need to be filled.
Next, a high-precision inkjet printer is used to print the "color" parts of the painting that need repair according to the information marked on the "damage map" onto an ultra-thin transparent polymer film. In addition, a layer of white film is added to increase color saturation and covered over the painting, making it appear to have the color "repaired" by artificial intelligence.
Since the covering film and the transparent varnish used for fixing can be removed with existing restoration materials without damaging the condition of the painting, it is advertised as being able to restore historical paintings in a "reversible" manner.
In proposing this restoration technique, Alex Kachkine pointed out that the restoration of many historical paintings is slow due to the high cost, time required, and the risk of failure. However, using Alex Kachkine's proposed technique, restorations that once took weeks, months, or even years can be drastically reduced to just a few hours, and paintings can even be easily restored to their pre-restoration state.
Alex Kachkine actually performed the restoration of a severely damaged 15th-century oil painting, using 57314 colors and repairing 5612 damages. It took only about 3.5 hours. However, if the restoration was done using traditional manual methods, the total time would probably be more than 66 times longer.
Alex Kachkine still emphasized that all restoration work still needs to be professionally evaluated and supervised by restorers with historical and artistic knowledge to ensure that the restoration results respect the original intentions of the artists and take into account artistic ethics and preservation principles.
