Loading...
> >

Remixing your art with AI

1…1517192021

Comments

  • DiomedeDiomede Posts: 15,128
    edited November 2023

    Question for the 'AI is not art' adherents.  Am I more or less of an artist if I use the following workflow instead of a workflow using a genesis figure in Daz Studio?

    (1) Modeled, rigged, and posed my own human figure in Carrara.

    (2) Created and conformed my own clothing and hair in Carrara for the figure in Step 1.

    (3) Used an AI program to detect the pose in step 2.

    (4) Used an AI program to detect the outlines in step 2.

    (5) Used an AI program to generate an image using the poses and outlines for a specified style.

     

    bb01 brash model.png
    1575 x 982 - 312K
    bb02 brash rigged and posed.png
    1546 x 941 - 301K
    bb04 openpose detect from image.png
    1830 x 993 - 250K
    bb07 canny fuction detect edges.png
    908 x 561 - 88K
    bb06 low res result.png
    512 x 768 - 584K
    Post edited by Diomede on
  • DiomedeDiomede Posts: 15,128
    edited November 2023

    double post

    Post edited by Diomede on
  • ArtiniArtini Posts: 9,331

    Thanks, Diomede.

    I am trying to find inspiration for my own creations,

    even if it is a long way to go.

    Just thinking about possibilities of images matching,

    that one see in Adobe Photoshop Elements 2024.

    Maybe such feature exists in the other image editing programs.

     

  • ArtiniArtini Posts: 9,331
    edited November 2023

    Prototype and Thor with AI background in Unity

    https://www.daz3d.com/prototype
    https://www.daz3d.com/cargo-platform-thor

    image

    Prototype01.jpg
    1920 x 1080 - 496K
    Post edited by Artini on
  • WendyLuvsCatzWendyLuvsCatz Posts: 38,056

    just a Genesis 9 render with a Deforum Stable Diffusion animation in the background

  • ArtiniArtini Posts: 9,331

    We got a new AI tool from Daz 3D:

    https://www.daz3d.com/face-transfer-2

     

  • ArtiniArtini Posts: 9,331

    Nice hair and cloth animation, Wendy.

     

  • DiomedeDiomede Posts: 15,128

    Artini said:

    We got a new AI tool from Daz 3D:

    https://www.daz3d.com/face-transfer-2

    Wow - well spotted, @Artini.  From the product description.  "Face Transfer 2 is an AI-powered system that from a photo transfers face shape, texture, and tone to create a look-a-like 3D version of family members, yourself, really whomever you want! It is compatible with Genesis 9 figures, employs higher-resolution texture maps, refines the image projection process, and improves dynamic geometry adjustments."

  • DiomedeDiomede Posts: 15,128

    Artini said:

    Nice hair and cloth animation, Wendy.

    +1

  • DiomedeDiomede Posts: 15,128

    Here is another example of an attempt by me to mix my custom models with AI, intending for comic style.  Everything in the base scene from which the following image was derived was modeled, rigged, etc by me, including the human figure.  See models in my post from November 23rd above.  I then used my base render output from Carrara in the Controlnet channels of Stable Diffusion, along with text prompts, to get the following.

  • WendyLuvsCatzWendyLuvsCatz Posts: 38,056

    Brash lookin' flash

  • ArtiniArtini Posts: 9,331

    Another nice image, Diomede, in the style, I also like.

    I have spent some time during the weekend experimenting with Unity,

    trying to blend Daz 3D assets with AI backgrounds, but still need more time.

     

  • ArtiniArtini Posts: 9,331

    Also spotted your workflow with AI, Diomede.

    Great explanation of the process.

     

  • ArtiniArtini Posts: 9,331
    edited November 2023

    Simple, but nice looking...

    image

    Rag11.jpg
    1024 x 1024 - 192K
    Post edited by Artini on
  • ArtiniArtini Posts: 9,331
    edited November 2023

    I like this image, too.

    image

    Rag16.jpg
    1024 x 1024 - 245K
    Post edited by Artini on
  • ArtiniArtini Posts: 9,331
    edited November 2023

    Sadness in the eyes?

    image

    Rag33.jpg
    1024 x 1024 - 195K
    Post edited by Artini on
  • ArtiniArtini Posts: 9,331
    edited November 2023

    Could not resist a cat image...

    image

    cat04.jpg
    1024 x 1024 - 163K
    Post edited by Artini on
  • ArtiniArtini Posts: 9,331
    edited November 2023

    Happy doggy...

    image

    dog16.jpg
    1024 x 1024 - 186K
    Post edited by Artini on
  • ArtiniArtini Posts: 9,331
    edited November 2023

    Checklist?

    image

    Princess18a.jpg
    1024 x 1024 - 415K
    Post edited by Artini on
  • DiomedeDiomede Posts: 15,128

    Gorgeous!  @Artini

  • ArtiniArtini Posts: 9,331

    Thanks for the comment, Diomede.

     

  • ArtiniArtini Posts: 9,331
    edited December 2023

    Has tried image 2 image from Daz Studio render of Julius 9

    image

    to make him look more like from comics

    image

    Julius901pic05.jpg
    1024 x 1024 - 193K
    Julius901pic05ai.jpg
    2048 x 2048 - 594K
    Post edited by Artini on
  • WendyLuvsCatzWendyLuvsCatz Posts: 38,056

    DAZ and Carrara render but the music is entirely AI

    In a land of tales and glory
    A traveller set on his quest
    Through valleys deep, and mountains hoary
    With a song deep within his chest


    Oh, the melodies he sang
    Echoed through the starry night
    With each strum of his lyre
    He brought joy and light

    In a time of knights and castles fair
    A rover wandered with an adventurous air
    With a heart full of valor, he sought his fate
    Through ancient forests and tales of late

    Medieval rover, oh, where will you roam
    Guided by stars, to places unknown
    With sword in hand and a spirit so free
    Medieval rover, your legend we'll see

    In a world of castles and knights
    A rover traveled by moonlight
    With tales of kings and wars, he'd sing
    A wandering soul with a golden string

    Oh, medieval rover of boundless dreams
    Roaming the lands, where no one redeems
    Through misty forests and rivers wide
    His legacy echoes, forever to guide

  • WendyLuvsCatzWendyLuvsCatz Posts: 38,056

    video so far

    the weapons scaled up slightly with figures hidden rendered with Filament then batch processed through Stable Diffusion

    sandwiched between background and iray rendered actors with screen blending

     

    a000.jpg
    1920 x 1080 - 77K
    00001-1161281830.png
    1920 x 1080 - 523K
  • WendyLuvsCatzWendyLuvsCatz Posts: 38,056

    breakdown

     

    finished video

  • DiomedeDiomede Posts: 15,128
    edited December 2023

    Great stuff, Artini and Wendy.  yes 

    Artini, Image2image sure looks like a success.  Hope that is the look you were going for.  Was the hand in the pocket something you prompted for, or a happy accident?

    Wendy.  What do you use for your music AI?  Thanks for the breakdown and finished.  Inspirational.  I am still having error conflicts trying to use SD extensions related to animation.  I've asked about it and am told a solution is to disable and then re-enable the extensions at each step.  Ugh.  Is that what you do?  Or are your animation extensions working without conflicts?

    Post edited by Diomede on
  • WendyLuvsCatzWendyLuvsCatz Posts: 38,056
    edited December 2023

    Diomede said:

    Great stuff, Artini and Wendy.  yes 

    Artini, Image2image sure looks like a success.  Hope that is the look you were going for.  Was the hand in the pocket something you prompted for, or a happy accident?

    Wendy.  What do you use for your music AI?  Thanks for the breakdown and finished.  Inspirational.  I am still having error conflicts trying to use SD extensions related to animation.  I've asked about it and am told a solution is to disable and then re-enable the extensions at each step.  Ugh.  Is that what you do?  Or are your animation extensions working without conflicts?

    the breakdown one used a local build through Visions of Chaos of Meta's MusicGen https://huggingface.co/spaces/facebook/MusicGen 

    another online version here https://waveformer.replicate.dev/

    Visions of Chaos which I use for Auto 1111 as well as most other things https://softology.pro/voc.htm besides working on my PC it has a useful disc space management utility for removing python stuff you are no longer using from the venv etc folders without having to go full uninstall chase down remove all the stuff as it scans and lets you check stuff you don't want anymore

    machine learning can rapidly fill up your C drive otherwise

    my main video in that last one is me playing VSTi instruments on my MIDI keyboard, not AI wink

    A1111 self contained or github repositories do not work at all anymore on my rig, my iGPU has hosed many machine learning programs, fortunately not VOC so far

    I will need to address this some day with hardware, maybe a usb adaptor for my second monitor or a visit to the repair shop to find why my 2080Ti display ports don't work

    Post edited by WendyLuvsCatz on
  • ArtiniArtini Posts: 9,331

    Thanks, Diomede.

    No, the hand in the pocket was just a happy accident,

    but it looks ok for me.

    Now I spend more time exploring the new Face Transfer 2 using images generated by AI

    as a reference photos.

    What a huge area to explore...

     

  • WendyLuvsCatzWendyLuvsCatz Posts: 38,056

    various Carrara rendered layers batch processed with Stable Diffusion image to image

     

    original combined layered Carrara video

  • ArtiniArtini Posts: 9,331

    Great, looks like you have a lot of fun making these videos, Wendy.

    Happy New Year and all the best for you.

     

Sign In or Register to comment.