Experiments with tape and glue

9BBB438C-B067-4A92-B419-704E6DDEBEA9.jpeg


I know the source image for this, and I know it ended with a trip through Graphite, but I can't remember a thing about what went on in between. :rolleyes:

EDIT: Maybe it was Decim8 in the middle?
 
Last edited:
Art Theft, Too

9BE73655-0AA7-4009-B8E1-1170A4D69BA7.jpeg

I play with Fragment & Tangent a lot, but don’t get results like rizole ’s in MC. I think I’ll experiment with straight photos (I’ll have to take some :D) This is half-way: a crop of my art-theft Mona Lisa back & forth in both apps.
 
Last edited:
Nice pictures but your input is what? Choosing a phrase and pressing a button, hardly original innit.
Ever been told, that's a good photo, you must have a good camera? As if the tool does the work and you do nothing. Aren't you asking a similar type of question here?
The input comes from recognising and judging the end image as a good one. Dream produces something different each time. Even using the same words and filter. And just like the majority of pictures I take, they are not very good. But I have enough of an eye to know they're not very good and they never reach the public. The words you pick for Dream matter too. I imagine the ai engine was trained on real images so the output bears a relation to the real world.
I used reptile skin for this. I threw away 4 others and kept this one.
1639249923167.jpg
 
The input comes from recognising and judging the end image as a good one.
I don't disagree, but its still an AI generated image from presumably a stock of images from goodness know where. Its art with very little human input, press the button, get a pretty picture. Or do you think its painting that picture just for you?
At least with apps like Prisma, that heavily rely on AI, you actually start with your own photo.

I always assumed MobiTog was MOBile PhoTOGraphy, not computer painting by numbers. My apologies if I've offended anyone but this is just a lazy ass app. Get out and take some photos please.
 
I think the ai has been trained on all the same biased data sets the rest of them have. Use the word skin and I bet you can guess what colour comes out. It also looks somewhat rough, almost masculine.
Use the word lips... Guess which gender tends to come out in terms of colour and fullness. Also the skin in the lips pictures is, of course, white, soft and smooth.
 
I think the ai has been trained on all the same biased data sets the rest of them have. Use the word skin and I bet you can guess what colour comes out. It also looks somewhat rough, almost masculine.
Use the word lips... Guess which gender tends to come out in terms of colour and fullness. Also the skin in the lips pictures is, of course, white, soft and smooth.
I think it's been trained on things like titled images as well. I got reasonably appropriate results from

The Cabinet of Dr. Caligari
Bladerunner
JMW Turner
Rain steam and speed
Impression, sunrise

I'm tempted to try some novel titles, including a few that haven't been made into movies.
 
I think it's been trained on things like titled images as well. I got reasonably appropriate results from

The Cabinet of Dr. Caligari
Bladerunner
JMW Turner
Rain steam and speed
Impression, sunrise

I'm tempted to try some novel titles, including a few that haven't been made into movies.
"He's not the messaih, he's a very naughty boy" brought back relevant results. not worth posting mind. I also tried different elements I could later mashup into something new but didn't find the results very useful. My fave is this one, HDR setting:
1639251997738.jpg

AI isn't very productive/useful yet in image production In my opinion. Needs a few more years.
 
Last edited:
Thanks for not posting the results :eek: I’m still recovering from (and avoiding eye contact with) rizole’s reptile skin
I didn’t keep the brains, I didn’t like them so much. I also did “once upon a midnight dreary” steampunk and I forgot which other theme I used. Since this is an Experiment thread:
SteamPunk
B0BEFF9C-0F30-4DDD-B0DF-E177EE197ECA.jpeg

Other
5E200A1F-BEB0-464D-B0B5-0FFC0D8777A4.jpeg

I like “other” better.
 
I don't disagree, but its still an AI generated image from presumably a stock of images from goodness know where. Its art with very little human input, press the button, get a pretty picture. Or do you think its painting that picture just for you?
At least with apps like Prisma, that heavily rely on AI, you actually start with your own photo.

I always assumed MobiTog was MOBile PhoTOGraphy, not computer painting by numbers. My apologies if I've offended anyone but this is just a lazy ass app. Get out and take some photos please.

Just avert your gaze, John. Have a cup of tea and a nice lie down. If necessary, think of England.
*This too shall pass*
 
The Sufi dancer Womble through AdLib

0E7ED5E6-3CC7-4883-9971-4C6F2D6524C2.jpeg
F096CD43-8D93-479A-8316-52830B31496D.jpeg

I like what AdLib’s mono does with the texture and the strong lines.
Meanwhile, Womble has died on me. I’ve deleted and reinstalled a couple of times with no effect. It loads up, in a manner of speaking, but only part of the UI is loaded, no option to get started.
Thought: I should check to see if I can use it on my phone.
 
Back
Top Bottom