• We have updated the guidelines regarding posting political content: please see the stickied thread on Website Issues.

Artificial Intelligence (A.I.) Art

As I'm terrible at art, and have never tried to draw hands, I'd like to ask: why are they so difficult to draw correctly?
 
The biggest problem I think is our brains interpret a 2D image as a 3D image, so that we can navigate life in three dimensions. We're then asking it to reinterpret that 3D understanding as 2D again, and that makes perspective difficult. Hands have quite a few different elements which curve, look very different seen from different angles, and it's very obvious from our experience of what hands do if they're not right.

I found I started seeing the world differently when I was drawing a lot. I would look at things and a new part of my brain would kick in and see how it actually looked from my relative angle, and how that image created the understanding of what the 3D object. You look at a hand and you're brain sees a square with five long bits protruding from it, but unless the hand is flat on the table seen from directly above, a square with five things coming from it isn't what you're actually seeing.

I feel I've got an intuitive sense of what I'm trying to say but haven't found the right words to say it.
 
I just tried a website with a ""What would you look like as a,,," function. You upload a photo and it morphs your face into a picture of a viking or roman emperor. What surprised me was that even with just 1 photo, from the front, it managed to give me some images in 3/4 profile which looks like I could have posed for them. It managed to analyse my face that well in just seconds.
 
Steven said:
why are they so difficult useful to draw incorrectly?
Click to expand...
To make us Natural Idiots believe that Artificial Intelligences actually have any limitations.

It's an overt shortcoming in order to lull the cattle into thinking they can spot the wolves through their pathetic camouflage.

These ridiculous suggestions are no less flawed than unquestioned acceptance of the illogical paradox that AI art always get humans' hands wrong.
 
Last edited:
I asked for a waterfall it looks pretty good. Surely there must be something wrong though, there usually is.
 

Attachments

  • IMG_3954.jpeg
    IMG_3954.jpeg
    406.7 KB · Views: 21
I`ve been playing around with a website of "What would you look like as a..." function. It is like seeing your previous lives.
Now it has me wondering what the ethics are of using those photos on dating websites, as I look pretty cool in some of them.
 
Xanatic* said:
I`ve been playing around with a website of "What would you look like as a..." function. It is like seeing your previous lives.
Now it has me wondering what the ethics are of using those photos on dating websites, as I look pretty cool in some of them.
Click to expand...
Ask it what you’d look like as a pic on a dating site then.
 
I used AI to imagine some of the most iconic mythical creatures — this is how it turned out

Ideas can come from anywhere and often when you least expect them. That is the case with this story, which was born out of a discussion about Burns Night and wondering what a “Haggis” animal might look like if one were to exist. This led to the idea of using AI to bring life to the mythical creatures of lore.

NyWA8afNk5EArGkCPErvYD-1200-80.gif


The first image I generated was of a Haggis, leaving it up to the AI to determine exactly what its appearance might be. In the end, it came up with a cross between a particularly wooly sheep and a pig.

StG2CFPgPDRH23jk7RtmLm-1200-80.jpg


Chupacabra

For this project, I turned to Leonardo, which uses its custom versions of the Stable Diffusion models as a base. The reason for using Leonardo was because of its Motion feature, which very quickly allowed me to create an AI video based on each image I generated.

https://www.tomsguide.com/ai/i-used...-mythical-creatures-this-is-how-it-turned-out

maximus otter
 
I tried some Cabin Pressure (The best radio sitcom) pictures. Benedict Cumberbatch did better than Roger Allam or John Finnemore.
 

Attachments

  • IMG_4260.png
    IMG_4260.png
    1.8 MB · Views: 9
  • IMG_4261.png
    IMG_4261.png
    1.5 MB · Views: 9
  • IMG_4262.png
    IMG_4262.png
    1.8 MB · Views: 14
Roger Allam ... yeeeees, I see the resemblance. Buty Finnemore? Nope.
Why do they look as if they're in a submarine? It doesn't look like an aircraft galley.
 
Stormkhan said:
Roger Allam ... yeeeees, I see the resemblance. Buty Finnemore? Nope.
Why do they look as if they're in a submarine? It doesn't look like an aircraft galley.
Click to expand...
It looks like Roger after a bad day. Maybe Ai doesn’t know the difference between an aeroplane and a submarine.
 
Stormkhan said:
Roger Allam ... yeeeees, I see the resemblance. Buty Finnemore? Nope.
Why do they look as if they're in a submarine? It doesn't look like an aircraft galley.
Click to expand...
Carolyn not so good. Herc better. I asked for 100 otters on the flight deck, it couldn’t manage that.
 

Attachments

  • IMG_4337.jpeg
    IMG_4337.jpeg
    100.3 KB · Views: 9
  • IMG_4338.jpeg
    IMG_4338.jpeg
    323.7 KB · Views: 9
  • IMG_4335.png
    IMG_4335.png
    1.9 MB · Views: 13
  • IMG_4336.png
    IMG_4336.png
    779.5 KB · Views: 10
There is a board game which I enjoy but they no longer sell it, so I can`t buy my own copy.
I decided to recreate it. In the old days I would probably need to commission an artist to draw all the different character portraits for me. Instead I used an AI to do it in 30 seconds. A bit scary.
 
I had a strange experience entirely of my own making when I was driving down to Wales during the King’s Coronation. Listening on the radio, for some reason, I pictured the whole scenario like an end of season episode of Bagpuss with the mice in the pews, Professor Yaffle as the Bish and a bottletop as the crown. Kept me entertained the whole journey but that’s just my head. And my imagination I have yet to deal with.
 
It’s not great at trying to give me a Victorian Policeman. It can’t handle the helmet at all.The fingers are interesting on the last one.
 

Attachments

  • IMG_4895.jpeg
    IMG_4895.jpeg
    393.5 KB · Views: 13
  • IMG_4896.jpeg
    IMG_4896.jpeg
    470.1 KB · Views: 13
  • IMG_4897.jpeg
    IMG_4897.jpeg
    438.4 KB · Views: 11
1000010163.jpg

So my son brought this to my attention. Apparently there's a whole world on social media of AI generated art. People share AI generated pictures on social media to get as many likes as possible. So popular images are combined to garner more likes. These are shared on some occasions by bots, so AI generating art based on popularity, many with a religious theme, with an element of danger, sometimes made out of cans, with likes generated by other bots. Leads to some truly bizarre images like this one which has religious elementd combined with a child and coffee.
 
You must log in or register to reply here.
Back
Top