Being born in the modern age, we often forget how different things were before us. People could only know what someone looked like based on an artist's subjective vision or a grainy black-and-white photo. Despite the efforts of the artists, paintings can rarely achieve the level of realism that pictures can. And thus, people often wonder how some of the most famous historical figures really looked with all of the subjectiveness removed from the image. Bas Uterwijk, a photographer from Amsterdam who has a background in computer graphics, 3D animation, and special effects, slowly but surely tries to provide the answer to this aching question.
He started in 2019. After experimenting with Billy the Kid's photograph and seeing positive results, Uterwijk decided to try and recreate someone who lived in an era where there were no photographs. He tried Napoleon, but didn't want to share the results because they were "nice, but not perfect." About a few months ago, he came back to Napoleon and tried it once again. "The software I work with is developing fast," he thought, so he figured trying it at a later time might bring better results. You can see how his "current" Napoleon turned out in the picture below.
Fayum Mummy Portraits
Here's what he told Bored Panda: "I think the human face hasn't changed dramatically over thousands of years and apart from hairstyles and makeup, people that lived long ago probably looked very much like us, but we are used to seeing them in the often distorted styles of ancient art forms that existed long before the invention of photography."
He started experimenting with creating images by using generative adversarial neural networks, because he thinks it will one day be the "successor to photography." "These 'Deep Learning' networks are trained with thousands of photographs of human faces and are able to create near-photorealistic people from scratch or fit uploaded faces in a 'Latent Space' of a total of everything the model has learned."
Artbreeder, Bas's program of choice, uses artificial intelligence and neural networks to create near-photorealistic images based on his input combined with what it knows about how faces look and how light behaves in photography. "I try to do everything as procedural as possible and let the AI do the most work, but sometimes I need to cheat a little in Photoshop because these GANs don't do clothes or classical hairstyles yet."
"Although I see my creations more as artistic impressions than as scientifically justified, in some cases, the results for me feel much closer to reality than most other methods that are used in reconstructions of people so far," Bas expressed. They say that a portrait artist's main goal isn't getting the little details right, but capturing the "essence" of the person's character. Do you think Bas has managed to achieve that? Tell us in the comments!