Using machine learning to recreate photorealistic portraits of Roman Emperors | Hacker News Reader

Using machine learning to recreate photorealistic portraits of Roman Emperors

voshart.com

356 pointselorant128 comments

This is undoubtedly cool, but I worry quite a lot about people who aren't educated about ML interpreting these "historical renditions" as a "ground truth forensic". Rendering fossilized animals and coloring black and white photographs is also cool, but when you start to render the bone structure and "caucasian-ness" of statues, it starts to veer into some uncomfortable racial and ethnographic territory.

What data and models were used to synthesize these predictions?

This is absolutely true. It seems like the author did extensive research looking for information about physical features of each emperor to try to produce an accurate rendition. If you look in the links at the right of the page, you can find pretty good documentation on the sources used. I think the link below would actually be a better one for HN.

https://medium.com/@voshart/photoreal-roman-emperor-project-...

>Artistic interpretations are, by their nature, more art than science but I’ve made an effort to cross-reference their appearance (hair, eyes, ethnicity etc.) to historical texts and coinage.

I have to agree with the grandparent, this seems like not very strong an underpinning for a project that clearly has the potential to be very controversial. There's no mention of accounting for the artist's potential implicit bias, the drift of ethnically-associated physiological features over the ~2000 years between the subjects and the data set that the process is based on, etc.

For a long time now, I've felt that the best practice for this sort of presentation isn't necessarily a static image, but an interactive experience that allows viewers to explore a range of reasonable outputs.

I agree with the grandparent as well to a degree. The statement "I have created photoreal portraits of Roman Emperors" is tenuous at best. I'm not intending to suggest that the creator of this work didn't spend a significant amount of time and effort or that the end result isn't very impressive. But I do think a disclaimer is warranted in that there's no guarantee that the resulting images actually look like the emperors. I also agree that some type of interactive visualization would be great. But I think these images are very cool as well :)

You mention black and white colorization and that was my first thought upon seeing this. The statues are already super “photo-realistic” so I’m guessing this was a pretty similar process.

On a completely different note, I am not sure what you’re getting at with the racial stuff. Like what are you suggesting? That perhaps some of the emperors weren’t totally Caucasian but rendering them as such from the best available evidence is somehow racist? In my humble opinion, trying to find racial undertones in every tangential thing is a distraction from the real conversations that we should be having. And there’s a real possibility of alienating people that are otherwise sympathetic to the cause.

> trying to find racial undertones in every tangential thing

In this case it isn't tangential but central, since the point is extrapolating physical appearance from incomplete data. It matters if the results predict inaccurate racial features because plenty of people unfortunately care about kind of thing.

And remember that laypeople read "researchers used machine learning" as "super smart people used advanced science to get this result so it's probably true."

But, isn't "extrapolating physical appearance from incomplete data" the entire point of this project?

It is, but the criticism is that it's easy to misunderstand and/or skip over the "incomplete" part, and get a wrong conclusion that the results are more accurate than they actually are.

My understanding of the original point is that ML models are trained on a data set. If you train the model on a data set that is only made up of white faces, the results will be whiter than if you train it on a data set that is made up solely of black faces. They said they'd like to know the makeup of the data set that the model is trained on to more fully understand how the results were produced.

But there was no attempt made to account for the incompleteness of the data in this particular known way.

I feel like the current climate of society shows that people who have such a simplistic view of science don't tend towards "smart people said this, so it must be true"....

They do, they just often can’t make a difference between snake oil salesman from a good scientist.

And that’s exactly the point.

These images were rendered with a "white" bias. Facial features is one thing, but skin color is clearly not conveyed by statues, yet a clear choice was made here.

You can't just claim this without suggesting what's actually wrong. What is wrong, and where are the good sources from which we can make a comparison? If you're just guessing then you're no better than this ML model.

Given the preponderance of blonde emperors, it would be reasonable to assume emperors in general were rather light-skinned.

What is blond, in a two-thousand-year-old context?

https://www.youtube.com/watch?v=gMqZR3pqMjg

It doesn't matter. Non-white people without steppe ancestry are pretty much uniformly dark-haired and don't have names akin to "subflavum" for traditional hair colors, because light yellow/gold hair just isn't a thing, regardless of what exactly they conceive of as "yellow/gold".

Aboriginal Australians have blonde hair arising independently.

Pretty sure Augustus wasnt aboriginal Australian, though (I see your point)

Yet these renderings were not made with that choice.

There are a fair few of them depicted as what I'd consider blonde. Augustus particularly.

Augustus was described as "subflavum" which is often translated as "blonde" but is more likely to mean "light brown".

Sadly this _is_ an area which is indeed full of bias.

Is that not what 'blonde' generally is in adults? An image search for 'blonde man' returns men I'd consider 'blonde' which hair which seems objectively light brown (except those with obviously bleached highlights.)

I wouldn't know.

The color of Augustus here[0] is definitely blonde for me, bordering on platinum, while what I meant is that it might as well have been darker, i.e. [1] which I would call "light chestnut".

[0] https://voshart.com/ROMAN-EMPEROR-PROJECT [1] https://external-content.duckduckgo.com/iu/?u=https%3A%2F%2F...

There are countless texts ( written at that time ) describing them.

This wasn't at the Jurassic age.

> That perhaps some of the emperors weren’t totally Caucasian but rendering them as such from the best available evidence is somehow racist?

These days, sadly, yes. The definition of racism has been so diluted in recent years that, as you implied, I fear it's often a distraction from actual racial injustice. Anyone ever hear of the boy who cried wolf?

Downvote away! :)

I don’t think gp was implying rendering as such to be racist, but was more concerned about people taking these renders as fact and using them to justify a prejudiced world view.

If someone is prone to accepting the idea that caucasians are superior to others they may use something as seemingly innocuous as this neat project to justify their positions and the lack of disclaimers towards the inaccuracies of certain aspects of these renderings may do a small harm societally.

On its own it’d be fine but in conjunction with other symbols and narratives it can become part of a larger, more sinister narrative.

People on a confirmation shopping spree have plenty of things to buy already. Historical accuracy is still a good aim.

> On a completely different note, I am not sure what you’re getting at with the racial stuff

What I am saying is that this sort of work can lead to people believing these images as true, i.e. "this is what Roman Emperors looked like". If you ask an ML system to solve a problem for which the inputs cannot possibly determine the answer (i.e. the color of a fossilized animal or the skin color from a person's statue), it will give you a highly biased answer.

I fear that some laypeople think of ML as "magic software that can do anything" without considering whether a problem is non-ambiguous given its inputs (e.g. something like object classification is less ambiguous than this problem).

I'm not saying "burn it to the ground" - I wanted to learn more about what data and model was used to generate these images so that I might understand better where potential sources of racial / ethnographic bias might come from.

> trying to find racial undertones in every tangential thing

If you read my comments history you'll see that I'm no critical race theorist, nor do I discuss everything in a racial context.

Out of curiosity, what "cause" are you talking about?

>"this is what Roman Emperors looked like"

The same can be said of the original advanced technology used to render a facsimile of the Roman Emperors, and in fact .. many stone sculptors have lost their prestige for the sake of this very same argument.

Since culture is a lie which persists only in the telling, this IS what the Roman Emperors look like, now. You shouldn't resist that, because there is no truth on this issue, anyway.

Is there anything in the HN guidelines about willfully dismissing or obscuring the truth? This comment makes me feel like there should be.

Saying "we'll never know anyway so stop asking" is abhorrent to me, and I imagine many on this forum. Not to mention it aims to end a discussion it doesn't understand.

Well, the sculpture is a fiction, and this AI generated image is a fiction - what's wrong with acknowledging that these fictions are all we have? Life is a collection of these facsimiles.

Acknowledging that is fine. Saying that makes further conversation irrelevant moves us in the wrong direction as a society.

Money is a fiction, history is a fiction, most scientific concepts eventually become fiction at a low enough level. And look at the world we've built on top of those fictions. It may not be perfect but it is undeniably meaningful.

How useful is it, then, to say .. "this is not accurate" .. well, there is no way to be sure. Ever. Unless .. you've got a time machine?

You'll never know with certainty, but you can certainly know more than you do now. But you won't if you don't pursue the truth, and it becomes a self fulfilling prophecy. As I said in another post, maybe it's possible to sequence DNA of the emperor or a collection of contemporary subjects, determine an approximate ethnic makeup, and revise the images based on that.

Or, we can just assume we are already correct because we'll never know anyway.

I don't think the parent comment was saying "stop asking" merely that "the answer is unobtainable, so any discussion about it is going to be grounded on that"

I agree. I think the problem very much starts with the sculptures themselves. It seems rather likely that the sculptors would have had an incentive to "improve" some of the not so attractive features of those very powerful men.

A realistic ML model could theoretically compensate for that by reintroducing a bit of variability based on the variability we see in today's population of that region.

Given the historical record, assuming the population remained the same is not even remotely warranted.

It certainly hasn't remained exactly the same, but there could be quite a bit of similarity left:

It is generally agreed that the invasions that followed for centuries after the fall of the Roman Empire did not significantly alter the local gene pool, because of the relatively small number of Germanics, or other migrants, compared to the large population of what constituted Roman Italy.

In 2008, Dutch geneticists determined that Italy is one of the last two remaining genetic islands in Europe, the other being Finland. This is due in part to the presence of the Alpine mountain chain which, over the centuries, has prevented large migration flows aimed at colonizing the Italian lands[1]

The migrants comprised war bands or tribes of 10,000 to 20,000 people,[5] but in the course of 100 years they numbered not more than 750,000 in total,[citation needed] compared to an average 40 million population of the Roman Empire at that time.[2]

[1] https://en.wikipedia.org/wiki/Genetic_history_of_Italy

[2] https://en.wikipedia.org/wiki/Migration_Period

That's Roman Italy - but what about North Africa and the Middle East, which were large parts of the Roman Empire?

On the attractiveness point, the author mentions that in the case there were multiple visual sources, the one which would likely be considered less attractive was chosen. (Under the assumption that it was probably more accurate since as you stated, there was likely incentive for flattery.)

Colors are can be subjective and somewhat politicized even when there's a written description. It's kind of like how every culture tends to color Jesus to be more similar to the local people, even though everyone knows he's from the middle east.

I might call someone blond with blue eyes, but someone else might describe the same person hair with light brown with green eyes. If we had third person choose which colors represented the descriptions, they might pick completely different colors based on what they thought blond or light brown meant.

We know what colors some statues originally had: https://art-sheep.com/the-true-colours-of-greek-and-roman-st...

Caligula looks to be on point based on the restoration of his original colors in that article.

To your second point, that reminded me that Japan's traffic signals use a different color "green" (looks more blueish) than the rest of the world and they have a common, named color that doesn't really map to blue or green in the rest of the world: https://en.wikipedia.org/wiki/Ao_(color)

The "racial stuff" seems relevant here. Sure, some people try and insert race into everything, but in this case it's a relevant, and interesting aspect of the problem domain.

My understanding is that the racial identity of the ancient Romans are actually a lot more ambiguous then most people realize, and the way the author attempts to reconcile that ambiguity has a huge impact on the image output here. And it's an interesting problem, Yann LeCun has discussed a bit about how to account for racial diversity via ML techniques (it's not just about the training data set!).

From what I recall, the reason for this racial ambiguity is because ancient Rome (and the entire Mediterranean region) was a cosmopolitan hotbed where many different ethnicities were mixing together all the time. On top of that, they didn't have the same kind of racial constructs we had, and so it's actually unclear what the races of various ancient Romans actually were.

Would be interesting to sequence their dna, and adjust these based on ethnic generic makeup

The way I like to explain it to people is that culture, ancestry, and language are 3 separate (but related) things. Knowing one doesn't necessarily tell you anything about the other two. In this case, it would tell you that the genetics are not sufficient to describe the categories of Roman ethnicity without additional justification that in this one particular case, they are.

I meant adjusting the images based on busts, not categories.

What does it mean to “render the bone structure of statues” ? A statue is already a three dimensional rendering. Besides Caucasian bone structure is supposed to include all of Europe, the Middle East and North Africa which is pretty much the bounds of the Roman Empire. I don’t remember any black or Asian Roman emperors so I have no idea who you think is being excluded here.

Not to mention that "bone structure" is not a fixed thing. The last 100 years or so, western skulls have seen a reduction in the size of jaw bones and sinuses, smaller mouths, narrower palates, etc. Apparently because we spend less time chewing on tough foods.

I commented about this above, but the race of various Roman emperors is not as clear as we think due to the cosmopolitan nature of the region, and different (lack of?) racial constructs. I believe Septimus Severus, for example, might have been black.

https://voshart.com/BIO the artist appears to be pretty socially conscious based on his other work.

Your comment frustrates me greatly because you are correct that people may internalize ML-generated images, and that's a real risk. ML already propagates racial bias within our prison system, as I'm sure you're aware. But Voshart's model isn't making decisions about employment or imprisonment, no, he's literally making art.

Art needs to be allowed uncomfortable racial and ethnographic territory. If not art, what do you think should occupy that space? Art is supposed to be the domain where we get to try out all these ideas and give the public exposure to what things like ML-generated images can be.

If this art made you feel something racial, try to do something positive with that. Maybe get involved solving some of the real life ML deployments that are already causing problems: https://www.technologyreview.com/2019/01/21/137783/algorithm...

But mostly, I hope when the author reads your comment, he doesn't feel discouraged that doing anything in public might risk his career. Or that someone thinking about publishing their own art doesn't look at it and go "hmm, too risky maybe, I better not". The chilling effects that comments like yours have is much more damaging than any racial bias you might be trying to raise awareness about.

I remember when the conscious thing to do was to uplift the art of the unseen people in our world. I wish we could go back to that.

Parent comment seemed like a discussion, not an attack. He started with, this is really cool. What is art for if not to provoke emotion and discussion?

It gets into confusing territory when we talk about who is being "silenced" in a simple conversation where folks disagree, but I think the best outcome is that we sell to understand each other's points first, and then share our own. I hope I've demonstrated that here.

I think these things are mostly for interest generation articles and possible exhibitions at museums. Artists have been sculpin these things since before the digital age, and while doing it with ML is cool and all, the results aren’t really more historically useful than what the sculptors did. But hey, if it generates interest I don’t see anything wrong with it either.

I wouldn’t be too worried about whether or not people view them as historical renditions. Most layman’s history “renditions” are mostly somewhat made up.

Your comment sounds a bit like obscurantism ? The classic "You shouldn't do this, it's bad because you're biased. Besides, it's impossible to not be biased, so don't bother". Even though there are documented descriptions of the physical aspect of these people.

https://medium.com/@voshart/appearance-of-the-principate-pt-...

Creator here to answer: I have tried to make it clear in interviews these are not ground-truth representations.

"each step towards realism is likely a step away from ground-truth." via https://www.smithsonianmag.com/smart-news/peer-past-these-ph... & https://twitter.com/dvoshart/status/1292107503543689217

I have not in my Roman Emperor write-up itself been as overt. I thought it might be a bit too obvious. My writeups make it clear when no information about skin tone is available.

I did include such a clause in a forensic facial reconstruction experiment: saying: "My gut instinct is that the results are simply easier to look at, that each step towards realism is a step away from ground-truth skull measurements." because it was an experiment where there are likely living relatives.

https://medium.com/forensic-vr/ai-forensic-facial-reconstruc...

Proof: https://twitter.com/dvoshart/status/1294977757756227585

Thanks Daniel for your response! I hope my GP comment didn't come across as hostile or discouraging to future artists who want to explore this kind of art. I want to be clear that I'm not offended by your work.

My discomfort stems more from my belief that a non-technical / layperson audience is much more likely to conflate "each step towards realism" with a "step towards ground-truth", the exact opposite to what you suggested. The writeups make clear when information is not available, but the artistic medium itself (a rendered still image showing only one possibility) does not showcase this.

> In response to Cocci’s findings, Voshart removed all mentions of the site and revised several portraits to better reflect their subjects’ probable complexions

You graciously made corrections for plausibly Nazi historical propaganda. However, not every artist/historian may elect to do this, out of scholarly pride or a personal politics or just laziness. The artist subsequently wields tremendous power over the images that show up on Google Image search, and their decisions may unintentionally contaminate future historical research by way of inducing subtle prejudices.

We live in an age of misinformation - and I think the information that must be protected most carefully is information of the past, specifically around people. DeepFakes of living celebrities are fairly easy to verify and disprove, but as soon as you go back ~5 decades or so, some people start to challenge whether certain events (Holocaust, Tiananmen Square Massacre) even actually happened.

Some ideas on how one might mitigate this:

- Provenance technology that shows the editing history (potentially including a MD5 of the dataset) of how an image was synthesized.

- A simple, unobtrusive watermark on the center of the image indicating that it is but one of N possibilities, where N gives a rough estimate of the entropy over the distribution. Sort of like the recycling triangle numbers.

- GIF form showing multiple possibilities given the data you have.

I think there is a grand misunderstanding around my use of machine learning. Your first two suggestions have an assumption that the images weren't heavily modified before going into Artbreeder which was usually the case. or modified in a circular manner, Artbreeder>Photoshop>Artbreeder>Photoshop any kind of MD5 is lost. I could show a progres of each but it is absolutely enormous. Potentially hundreds of decision trees.

As for GAN bias: once in Artbreeder the skin tone is entirely my choosing. If there is a bias in the depiction it is my own bias. There might be a slight bias towards a white, brow-haired female but is is more of a gravity that can be avoided with a bit of skill. There is also a challenge in hair that is -for lack of a better explanation- hair made with human intervention (braids, tall wigs, cornrows, dreads etc).

AS for your last suggestion: I've already been thinking of making videos that show a transition of skin ton of lightest plausible to darkest plausible. Tiberius for example I have a very compelling version as a pale-skinned white dude or a dark-skinned person who looks indian/pakistani. Done in paintover/colorization in Photoshop --directly on the bust- where proportions match.

In the second version of the poster I actually challenged myself to do at least do a quick depiction of each emperor with dark skin. Version 1 my approach was to assuming an average based on place of birth. This time I experimented with assume-darkest-features possible simply to test my assumptions and question my biases. Although quick and not robust, I think it was of value.

What is uncomfortable regarding making an educated guess as to the race or complexion of a historical figure? I think that is reqlly sad.

> What data and models were used to synthesize these predictions?

You have to ask the creator of ArtBreeder (https://www.artbreeder.com/about) which apparently uses BigGAN, StyleGAN, etc., trained on different datasets to offer these options:

https://i.imgur.com/5jTNgXl.jpg

That being said, there is a heavy manual touch (based on "Photoshop and historical references"). See for instance how different Septimus can appear compared to the webpage:

https://twitter.com/dvoshart/status/1290004339646517250

This is essentially the same issue with things like artists conception of dinosaurs. Lots of people probably have strong mental images of what color a T-Rex is, despite scientific evidence being very scant. Maybe "artist's conception" is a good term to put next to these sort of ML-generated images?

Same technological challenges, but the rendering of past historical figures' ethnographic features has much higher social stakes and is a considerably more politically sensitive issue than how a T-Rex is rendered.

After your comment I was expecting to see all pale blonde pictures, but there was quite a variety in skin tone, hair color, etc. I remember reading that the Roman empire was diverse because of their history of conquering, but the extent of the diversity here really surprised me.

...Or were you saying that this is more diverse than is to be expected?

Always good to keep in mind with these kind of ML reproductions: https://twitter.com/Chicken3gg/status/1274314622447820801/ph...

Historical writings. It is documented Caligula was a redhead.

Unlike other people I find the portraits surprisingly diverse, I expected to see very white faces considering most "realistic" representations are from white actors and white statues erase features. Not a perfect project but interesting.

It would be interesting to compare the colourisations to where any of the emperors are known to have originated and whether they came from Northern or Southern Italian ancestry. A blued-eyed blonde from Northern Italy would be quite feasible.

Also, later on many were not Italian at all.

Yeah a lot of those guys look pretty swarthy (e.g. like me.) I can't say I expected that.

Caligula looks as I expected him though.

It's worth noting that classical statuary was typically painted and not left as blank white marble (the practice has persisted in some Catholic statuary). I doubt that any remaining pigment on statues (which might be recoverable through microscopy) was incorporated into the models for creating these images, but it would have been cool had it been.

For more on the painting of ancient statues, see, e.g., https://www.newyorker.com/magazine/2018/10/29/the-myth-of-wh...

Roman emperors came from all over the empire, so the diversity is to be expected (e.g. Heliogabalus was from modern Syria, and there's noticeably Philip the Arab)

What is odd here is that it's actually likely overdone, as the "blondies" in roman times were unusual enough that they were explicitly noted as such, e.g. Lucius Cornelius Sulla Felix was noted as "his golden head of hair gave him a singular appearance".

We have plenty of colored images from roman times through mosaics.

A lot of people seem to be taking the term 'photorealistic' a little bit too literally, when we say video games have photorealistic graphics, people don't assume they are literally photos of things.

These are just 3d models of statues created to look as lifelike as a photo. Nothing more nothing less.

They're no more accurate or not than the original statues. Unless we were to go back in time and physically see these people there's really no way to tell either way.

It's just an interesting interpretation generated by some algorithms and a large sampling of data. It is what it is.

> They're no more accurate or not than the original statues. Unless we were to go back in time and physically see these people there's really no way to tell either way.

That's why the author uses multiple historical descriptions of the looks of each emperor. So he can make the pictures more accurate than the original statues.

He also uses paintings and other statues for comparison.

This is fantastic. Some of the reconstructions look convincingly italian. There is a slight bias towards making their faces more square than they are in statues (and probably RL too), e.g. Caligula.

It's interesting to think that these emperors would find the obsession with exact skin tones in the comments here rather weird.

Also, with great power comes a brutal way to die.

Fairly regularly at the hands of your own bodyguards, going by some of the later entries.

What’s the consensus on whether Roman statues were intended to be realistic, or perhaps were idealizations, or even whether the statue makers had the ability to truly capture realistic features rather than work to a “type”?

“Lely's painting style was, as was usual at the time, intended to flatter the sitter. Royalty in particular expected portraits to show them in the best possible light, if not to be outright fanciful. Lely's painting of Charles II shows what was expected of a painting of a head of state in the 17th century. It emphasizes the shapely royal calves - a prized fashion feature at that time.”

https://www.phrases.org.uk/meanings/warts-and-all.html

There's not one Roman style and there were wild variations across different eras. I'll skip the archaic and early Republican stuff because it's either too rough or derivative to be called specifically « Roman ».

Late Republican busts are assumed to be very realistic. My understanding is that they're largely derived from mortuary masks made for the cult of ancestors that was prevalent at the time. It's obviously impossible to say how accurate they were without direct comparison to remains, but busts from that era show aging and « ugly » traits prominently.

The Imperial era with its state-sponsored cult of personality started the trend towards idealism. Defects were smoothed out, people represented at their best and efforts were made to have them impersonate attitudes and virtues. The busts of the Pax Romana remain highly individualized and it's easy to tell a character from another, even though they always looks somewhat too good.

It started getting weird with the Third Century Crisis, the Empire breaking up and the Tetrarchy. Portraiture became abstracted to the point it was more about representing the office that the person occupying it. Look up the sculpture of the Tetrarchs that's in Venice or the Colossus of Constantine, they're well known examples.

Being a good fit for Christianity, that's the style that continued into the Middle Ages, and the reason why people assume that artists became terrible in « the Dark Ages ».

> What’s the consensus on whether Roman statues were intended to be realistic

Would be a pity when people don't recognize you as the emperor when for some reason you are disconnected from your usual chain of command. Could those depictions have played a role that we would today describe as biometric fallback authentication? In that case they wouldn't want to overdo it with the idealization.

Also, looking at it from a completely different angle: many Roman busts simply don't look very flattering at all, and to me those of emperors don't seem like a big exception.

Noting down a weird phenomenon: I'm a huge admirer of Marcus Aurelius. His Meditations is sitting on the night stand next to me as I type this. But seeing him rendered this way somehow diminished the almost super-human image of him I had built up in my mind. He's just a man here.

you didn't get shit of his notes if you thought of him as a super-human

As you probably know, Marcus Aurelius would not have wanted you to think of him as a super-human! He writes a lot about his fallibility and his attempt to live a stoic life despite his failures.

In my mind, that's what makes him "super-human" -- his ability to reflect upon himself is a skill that most average human beings don't possess.

A very modest looking man though which is exactly how he would have liked to look.

I share your study of Marcus Aurelius’ writings (diary style for his own benefit, happy accident that his writings survive to modern times).

He believed in Summum Bonum, or ‘do the right thing.’

These days we mostly make statues of football players and coaches (both real football and metric football). I guess these are as close as we get to emperors.

But despite modern techniques for capturing a 3D image, they rarely look like the living counterparts. Nick Sabens statue looks more like Andy Griffith. Gary Patterson’s statue flatters by shaving off several pounds. Does anyone think the Rocky Balboa statue looks like Stallone?

Is there any reason to think that the Roman emperors statues are any more true to life? Even if I was the most talented sculpture in the world, I know that if I was commissioned to do a statue of a still-living Caligula I wouldn’t include any warts.

Seems like a good experiment for this tech: take some statues of people with known photographs, and see what imagery comes out.

> Does anyone think the Rocky Balboa statue looks like Stallone?

(This is the only one of those I've seen in person.)

I do actually. It's obviously not a perfect laser scan of his head, but I think it's quite clearly a depiction of the man, including the characteristic harsh lopsided nature of his face (which apparently is the result of nerve damage he sustained during birth.) That aspect of the statue is perhaps underplayed a bit, maybe out of respect, but I do think it's at least a recognizable reasonable depiction of him.

This is super cool, but I have noticed blue eyes and hair, as that gene was extremely rare at that time, what kind of referees have you used?

And one more thing I want to add, each time I see coloring of WW1 or WW2 or this one, suddenly that history becomes not so distant and those people do not look so technologically backward. On that note somehow war atrocities in my mind become unreal...

According to one Quora answer [1], 10% of Italians have naturally blond hair and 30% of them have blue eyes. I don't see why Italians at that time should look much different to Italians today.

[1] - https://www.quora.com/Are-blonde-hair-and-blue-eyes-quite-co...

Because it's historically ignorant? IANA ethnographer, but I expect that until the collapse of antiquity/migration period, only a vanishingly small part of the population did not share the most prevalent phænotype (dark hair etc.) in the cis-alpine part of the peninsula. I distinctly remember from school that blonde natural hair wigs were a prized possession among Latin female nobility because it was so rare to see and so difficult to obtain because of distance.

Maybe because in the several centuries in between some exchanges (mixed marriages) with the rest of Europe, including the northern countries where blue eyes and blond hair is more common took place?

Here's what's said about Caligula[1]:

> hair and beard worn long for a time after death of his sister.

Humans are curious beings. If my memory serves right his sister died AFTER Caligula went nuts so at this point in his life he had already gone murderous after recovering from an almost fatal sickness (or poisoned) but when his sister died of fever his murderous psycho heart must have felt so broken (or perhaps felt a bit nihilist) he grew full beard. Maybe there's a human history of letting beards grow when shit happens.

[1] https://medium.com/@voshart/appearance-of-the-principate-pt-...

growing a beard (and hair) to honour the dead is still practiced in southern greece at least

In college we had a professor who went over some of the things these guys did in their spare time. This article sums it up a bit. Really helps to put a face to the name... =P

https://io9.gizmodo.com/the-11-most-depraved-things-the-roma...

Vespasian looks uncomfortably like Lyndon Johnson.

I bet those emperors would be quite pleased to know that millions see their faces two thousand years later.

I find Tiberius, Caligula, Claudius and Nero unconvincing, "too normal" to fit the troubled characters that I first knew of watching I, Claudius.

Maybe I'm biased, but I feeel the difference even in the images that the reconstructions are based of.

Augustus seems very well done.

Cinematic portrayals can color our mental images. My mental image of Cleopatra is Elizabeth Taylor, as ahistorical as that might be. The oddest along these lines in my head is that I always picture J. D. Salinger as James Earl Jones, thanks to Jones playing a character in Field of Dreams who, in the novel, was Salinger.

Check the article on how each emperor was done: https://medium.com/@voshart/appearance-of-the-principate-pt-...

I also thought Nero looked weird, but after seeing the other statues, and the historical descriptions, I think it's probably quite good!

I don't know, when I saw Caligula, I thought "Wow, that's exactly how I imagined he would look", like he had that same immature-yet-assholish look of Joffrey in Game of Thrones.

The generated face seems a little wider. Also the eyes of all of them seem very calm. I expected more on the wild side.

Wasn't Gaius Julius Caesar [1] the first emperor? How come he is missing?

[1] https://en.wikipedia.org/wiki/Julius_Caesar

Caesar died before the Roman empire was established. His adopted son Augustus recreated the republic into an empire and is the first emperor.

The image of the emperors on the website has the header "Roman Emperors - The Principate"

It begins with Augustus: https://en.wikipedia.org/wiki/Principate

I don't know, Gordian II's bust looks more like Pete Townsend than the ML version would indicate. Maybe it's the eye color.

Most of the later emperors are listed there as "killed". Why be an emperor if you have an 80% chance of being murdered?

Because for a time, you get to be emperor of Rome. We still know their names, and their faces, thousands of years past. 80% seems decent odds for that kind of immortal fame.

What they knew at the time is that they would be gods.

They also knew that they would become almost unimaginably rich.

And compared to their successors in the title/office of pontifex maximus in certain later centuries, their annualized mortality rate wasn't even that bad.

Thing is, if you're the type that has the chance to become an emperor, you're "in the game anyway". I you don't become the emperor, you're still likely to be killed by the emperor or the potential emperor to remove a potential threat, just without the benefits of being the emperor and without the resources to try and protect yourself that being an emperor gives you.

Yeah, I find it amusing/unsettling given the number of emperors who have their death notice something like "Killed by his own troops in favor of Next Emperor", then you go to the next emperor and it says "Killed by his own troops in favor of Next Next Emperor".

I hope you're gonna do the Dominate too. Would love to get a look at Diocletian or Constantine.

Aurelius looks odd, most of his statues have him with much curlier hair.

Marcus Aurelius looks like a chill dude. Tallest to shortest?

I find it interesting how much some of them resemble current famous people (for example, Augustus who resembles Putin with more hair). I wonder how much of that has to do with celebrities possibly being more represented in the training data?

don't be fooled: this is a photorealistic rendering of statues and coins. We have pretty precise images of Augustus, except for colors.

https://en.wikipedia.org/wiki/Augustus#/media/File:Statue-Au...

> We have pretty precise images of Augustus, except for colors.

What? We have many statues of Augustus, but they are known not to be modeled on Augustus. Those are images of some other random guy.

what I mean is: it's not some ml software that came up with a putinesque visage, it's just a recoloring of the statue, which already had that nose, oval cheekbones etc.

Too symmetric features in most cases to be true...imho

Yeah - the statues are not necessarily accurate depictions to begin with.

You would probably expect the artist to idealise the features to a degree for their health's sake. A Christiano Ronaldo statue would not go down well with any that had delusions of being a deity.

As the creator I must say there is a lot of misunderstanding in all the comments above but people are making many valid points. On a side-note you might be interested to see that I have tried to reconstruct that horrible Ronaldo statue. https://twitter.com/dvoshart/status/1287888497286750208

There's a ton of memes where people have skewed photos to look like the statue that are hideously amusing.

The statue should have looked like what you produced: I can only guess bronze was hard to work with or redo when it goes horribly wrong.

Not quite. Gordian II, Diduis Julianus, Philip The Arab. They do have clear eye-asymmetry that is quite common today. Maybe it was part of our evolution?

Is it just me or does Augustus look a lot like Vladimir Putin?

I see it too; I think they both share a weak chin, long cheeks, and light hair/eyes.

Interesting that Nero and Caligula resemble, more than anything else, frat boys.

It’s weird that no one in this thread is jumping onto this comment to point out how stereotyping the racial features of the frat boy stereotype is an unfair racially discriminating generalization /sarcasm.

They were young, well nourished, and not hugely into keeping fit.

> frat boys

Isn't this a hateful term?

Reply on news.ycombinator.com