8 ms·
AI Gahuku: AI Generator Will Turn Your Photos into Renaissance Paintings
- cwoolfe 7y agoGreat work! Although it is interesting to me how it doesn't work very well for non-white faces. Which totally makes sense given that renaissance art training data reflects the time and place in which it was created.
- dylan604 7y agoFrom someone that is not an ML/AI evangelist, how is this great work? None of the images I've seen remotely look good. I have yet to see something that a real life painter not suffering from a stroke would be willing to have released.
- imhoguy 7y agoIt is not me! To guillotine! :)
- smitty1e 7y agoThis is the sort of great work that invites the question of whether "can" == "should". It's awesome that the work of great artists can be reduced to an algorithm. That effort is its own kind of art, and will see application (e.g. restoration work) far and wide. It could help with tutorials for students to get into these older styles. Despite all of the intermediate goodness, I still want the no-kidding product I buy to have had some actual human imperfection and idiosyncrasy injected. As explored by Rick Beato in the musical context here => https://www.youtube.com/watch?v=L-8EbHkc8tc https://www.youtube.com/watch?v=L-8EbHkc8tc
- new2628 7y agoI think you give these projects too much credit. They do not reduce the work of great artists to an algorithm and they do not turn user's photos into renaissance paintings. These projects are technically impressive and interesting, but art is not going anywhere.
- smitty1e 7y agoThat would appear to be the position of the moderation.
- superfamicom 7y agoTips for best results: no teeth, no beards, no dark skin, looking directly at the camera tends to make one of the eyes perfect and the other wonky.
- mensetmanusman 7y agoTips for hilarious results: lots of teeth, lots of beard
- basch 7y agoThis worked pretty well as a test picture. https://upload.wikimedia.org/wikipedia/commons/7/7d/Glynn_S._Lunney.jpg https://upload.wikimedia.org/wikipedia/commons/7/7d/Glynn_S.... OR, REO look good. RE1 turned him into David Tomlinson.
- ipsum2 7y agoDoes anyone have an example to look at? The ones in the header are small and darkened. Edit: here's Trump and Obama, using their official Whitehouse portraits: https://imgur.com/a/iWeX9tl https://imgur.com/a/iWeX9tl. Pretty funny results.
- basch 7y agoTry this. It's interactive so after you submit, you can choose your style. https://upload.wikimedia.org/wikipedia/commons/7/7d/Glynn_S._Lunney.jpg https://upload.wikimedia.org/wikipedia/commons/7/7d/Glynn_S....
- ipsum2 7y agoThanks - I didn't want to contribute to HN's hug of death, doing inference is probably very expensive for them.
- wenc 7y agoI wonder about that? In most cases once an algorithm is trained, running inference is just a function evaluation, which is usually computationally inexpensive. (the training is the heavy compute part)
- notduncansmith 7y agoOne of those exposes a pretty clear bias in Renaissance painting subjects.
- mattigames 7y agoYeah, it doesn't do blonde hair.
- dylan604 7y agoWhich is odd because didn't everyone wear powdered wigs? Seems like everyone would have a blonde wig (which should make things easier).
- hobs 7y agoService seems to be overwhelmed.
- DeathArrow 7y agoThis is what I've got using chimp pics: https://imgur.com/a/hkG0rJg https://imgur.com/a/hkG0rJg https://imgur.com/a/YZ3oWUm https://imgur.com/a/YZ3oWUm
- retSava 7y agoHoly smokes - the second one was unnerving to say the least!
- imhoguy 7y agoThat is creepy and genious. So who is the author of the artwork now? AI?
- qchris 7y agoThere's actually been some litigation around this topic recently. The general consensus seems to be that authorship gets assigned to closest version of the person who initiated the creation process of the resulting artwork. So, in this case, I believe it would be the person that selected and uploaded the photo to the service would have authorship--the actual process is considered to be more of a blackbox tool being used by that person. A (way more) sophisticated digital paintbrush, if you will.
- dylan604 7y agoWouldn't this be similar to using Photoshop or any other image manipulation tool? Adobe does not get ownership of output of their software. Why would we think some random website offering image manipulation would be different? As you say, the software is just a tool in the image creation process.
- StrangeDoctor 7y agowhat gets really messy is the training data. Renaissance paintings are (hopefully?) all public domain, but what would be the case if I used living artists' work, or the collection of disney cartoons?
- samcheng 7y agoSomeone should do this with the "Getty Museum Challenge" photos that are circulating social media right now!
- babo 7y agoIt clearly shows a strong bias, practically turns a woman into a man.
- drcode 7y agoIt really captured the personality of our president well: https://firebasestorage.googleapis.com/v0/b/ai-gahaku.appspot.com/o/twitter-images%2F035177eb-e209-42de-ae0d-88c7d9294c3b.jpg?alt=media&token=34b5c7b7-73f1-4814-a501-3abb681135c7 https://firebasestorage.googleapis.com/v0/b/ai-gahaku.appspo...
- kristopolous 7y agoGreat work, calling it The Hapsburger would be amazing.
- amiantos 7y agoMaybe this should be titled "nightmare generator" because most of the pictures I have tried, especially of my wife, have ended up with very frightening distortions. Might be cooler to pair this with cubist paintings instead, so the facial defects seem more like features.
- jw887c 7y agoYeah this doesn't work for non-white folk. Just makes us look like white folk
- amrrs 7y agoThis is mentioned in their site. Seems they've identified it. >Currently, we are confirming that the output of the AI artist has been biased. We hope to use a wide variety of learning data and increase the diversity of output in the future.
- johnzim 7y agoThat's almost for the best. The representation of non-European ethnic facial features / skin tones in renaissance art appeared to mostly be one of two groupings: 1) Accurate, naturalistic portrayals that almost certainly had an actual human sitting and; 2) Color choices / features that are to humans of other skin tones, what that picture of a cat in the still life painting is to our feline friends: https://twitter.com/chelsesaurusrex/status/995127939585851392 https://twitter.com/chelsesaurusrex/status/99512793958585139... There are some really interesting examples in this article (although explicitly African): https://thewalters.org/wp-content/uploads/revealing-the-african-presence-in-renaissance-europe.pdf https://thewalters.org/wp-content/uploads/revealing-the-afri...
- ggggtez 7y agoIt's not really kosher to suggest that it's "for the best" that it doesn't work well for other people. :) It's not like paintings of darkskinned people don't exist for the AI to learn from, even if you don't like the style that was used at that time.
- ramblerman 7y agoIt seems like the authors are truly limited by the data here. I mean, if someone was to do a similar project of Chinese Qing dynasty portraits, you would expect a bias to Asian faces. We have to be so conscious of bias in AI, yet in this case I wonder what the solution (if there is one) would be, given you genuinely have a biased data set to begin with.
- hkmurakami 7y agoAh title is misspelled, page says “Gahaku”, which is a slightly archaic term in Japanese for artist/Painter with a connotation of being a highly skilled and respected master of the craft. There’s also a sarcastic net slag meaning to the term as well but the creators are probably not using the term with that intent.
- tomc1985 7y agoCan they not use the word "masterpiece" on the site? A masterpiece is supposed to be superlative. You can't call every single output a "masterpiece"
- 4weekoldroses 7y agoAs a representational painter, I've been waiting for someone to do a write-up on how/why these 'AI' paintings aren't going to be equivalent to human paintings until 'AI' is itself equivalent to humans. Unfortunately, I haven't seen one yet. Maybe I'll have to do it myself.
- tomc1985 7y agoIMO most 'generative art' is about as creative as a markov chain text, just with images as the input instead of words. It'll be a while before anyone can assert any sort of equivalence to human creativity and have it taken seriously
- BubRoss 7y agoYeah, the internet trend of turning everything into hyperbole squared is obnoxious. These look like blurry insane asylum art projects. I don't think any reasonable person is going to look at this stuff and call it a masterpiece.
- leogout 7y agoI wear glasses. I've tried with many photos of me with glasses and it always looks like it can't handle it. Do you think it is because it was trained with paintings at a time when glasses weren't common ?
- mrb 7y agoCombine it with https://www.thispersondoesnotexist.com/ https://www.thispersondoesnotexist.com/ and you've got a generator of completely fictional renaissance paintings. Here is one example, quite good: https://imgur.com/a/9F7xoC0 https://imgur.com/a/9F7xoC0
- secondaryditto 7y agodid most of the subjects in renaissance paintings actually exist, or were they just imagined people/faces?
- ElonsMosque 7y agoI mean I’m aware there probably weren’t many black people during the renaissance. But it would be even cooler if this worked for people with different shades?
- kumarm 7y agoNice. Worked well for me for several photos. Is this model open source?
- wardnath 7y agoMost mashups remove my beard, more sample cases would be great.
- LudwigNagasena 7y agoQuite ironic to name an AI that can only paint white people with a Japanese word.
- masonic 7y agoWe've already seen the beta version: https://www.bbc.com/news/world-europe-19349921 https://www.bbc.com/news/world-europe-19349921
- lerie1982 7y agoWould like to try it but, "I’m sorry but we are having a lot of traffic at the moment. Please try again later." There also doesn't seem to be a "retry" button, unless I'm missing something, I have to keep selecting and uploading my image.
- koushikn 7y agoArt Transfer from Google works similarly: https://artsandculture.google.com/camera/art-transfer https://artsandculture.google.com/camera/art-transfer
- etaioinshrdlu 7y agoNot really though. That google demo looks like a single-pass feed-forward neural network trained to perform style transfer. The textures and colors get replaced while overall content is stayed the same. This project seems to work by finding your image in the latent space of a GAN model, and then re-synthesizing a new image from that vector. It's more like generating a whole new image which is targeting the overall look of an existing image, while jointly having optimized the generated image to look like it comes from a set of renaissance art. Edit: on second thought, this tool might be running too quickly to be doing optimization to find an image in latent space. It might just be fancy vanilla style transfer done nicely. Hard to tell.
- xivzgrev 7y agoOH GOD PUT IT OUT WITH FIRE!! It’s a cool tool but added about 200 pounds to my face. P0 roadmap feature request: flatter the subject, like Instagram does.
- Daub 7y agoAs a painter, I’m genuinely impressed with the results. However, the system is not capable of taking much initiative. It will reject anything other than a full-face portrait. Of all genres, portraiture is the most convention-driven. For example, 90% of most portraits are three-quarter view, with side lighting. Effectively, there are only a very limited set of solutions to the problem. Applying the same approach to a landscape painting would be a different class of problem altogether.