Karen X. Cheng Profile picture
Aug 30, 2022 9 tweets 5 min read Read on X
1/ Using AI to generate fashion

After a bunch of experimentation I finally got DALL-E to work for video by combining it with a few other AI tools

See below for my workflow -

#dalle2 #dalle #AIart #ArtificialIntelligence #digitalfashion #virtualfashion
2/ First I used DALL-E to generate outfits. I did this by erasing parts of my existing outfit and inpainting over it

Btw when I erased the entire outfit, the results didn't look as good. By keeping parts of the original, DALL-E was able to better match color and lighting
3/ But here’s the challenge. DALL-E works great for individual pictures, but it’s not designed for video. It won’t give you consistency from frame to frame.

Here was my early experiment. See, no consistency between frames
4/ It’s generating completely different outfits for every frame. But I want the same outfit to persist for several frames, which DALL-E currently can’t do
5/ After a bunch of experimentation, I discovered a program called #EbSynth by @scrtwpns

It's intended for painting style transfers, but I wondered if it could work for clothes...

Demo video by EbSynth
6/ And it turns out, it DOES work for clothes!

It's not perfect, and if you look closely there are lots of artifacts, but it was good enough for me for this project
7/ Finally I ran the video through DAIN which smoothly blends from outfit to outfit

It had the added bonus of giving my video artificial slow mo
8/ Here’s another example of a video I did using DAIN  with @jperldev a while back

@jperldev 9/ Shoutout to @paultrillo - he's been using DALL-E in fascinating ways and he inspired me to start experimenting with trying to get DALL-E to work for video

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Karen X. Cheng

Karen X. Cheng Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @karenxcheng

Apr 10, 2023
1/ Filmmaker's guide to GEN-2

Generate videos with nothing but words
Access to GEN-2 will start rolling out this week @runwayml

I put together this guide to help with your experiments for when you play with it
#gen2
@runwayml 2/ Filmmaker's guide to GEN-2

It gets interesting when you add movement

It works best with simple movements btw, and struggles with more complex or intricate movements
@runwayml 3/ A peek behind the process

(This is beta UI, the final UI will change)

Currently it takes about 2 minutes to generate a 4 second video clip
Read 8 tweets
Feb 20, 2023
Some of my early experiments with @runwayml GEN-1

It’s still early days of AI video - this tech will only get better. A whole new generation of filmmakers is gonna be able to make whatever they want on zero budget

See below for my process
@runwayml Early access to GEN-1 is available via waitlist here: research.runwayml.com/gen1

I'll be sharing more tutorials once access opens up to more people
@runwayml Shoutout to @DonAllenIII who majorly inspired me with his early GEN-1 tests.

He taught me the workflow of prompting with a MidJourney image - that seems to get much better results than prompting with text

Thanks to @notiansans for troubleshooting help!
Read 5 tweets
Feb 7, 2023
Microsoft just announced its new Bing with @OpenAI integration - using a model more powerful than ChatGPT

I’m on campus right now with access. Tweet reply what you’d like to test and I’ll try your query -
Access is currently available via waitlist
I’m told it will roll out to millions of people within a few weeks

bing.com/new
Tip: try things that search engines today aren’t good at

Instead, treat it like a personal assistant, that has intelligence and can handle tasks and basic calculations

Example: “Will the IKEA Klippan loveseat fit into my 2019 Honda Odyssey if I fold down the seats?”
Read 9 tweets
Dec 20, 2022
Using AI for design inspiration

We used Stable Diffusion Depth to Image to get the consistency - collab with @justLV

See below for our process

#ArtificialIntelligence #stablediffusion #interiordesign
@justLV Any image will work for this - you could use pictures of your real furniture too

But we thought it'd be fun to use dollhouse miniatures since they'd be easier to manipulate
Read 6 tweets
Dec 5, 2022
NeRF update: Dollyzoom is now possible using @LumaLabsAI
I shot this on my phone. NeRF is gonna empower so many people to get cinematic level shots
Tutorial below -

-
#NeRF #NeuralRadianceFields #artificialintelligence #LumaAI
2/ First, stand still. Have a friend take a quick video of you
Capture the area in front of you and behind you
3/ Here's the raw footage btw for anyone curious
(Feel free to download it so you can follow along)
Read 7 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(