The walls between the real world and what AI can generate continue to dissolve so let’s get into this week’s stories.
Welcome to WednesdAI – Pixel Dreams’ weekly update with top stories from the rapidly evolving world of Artificial Intelligence.
Subscribe to the brand new WednesdAI YouTube channel!
Real Time Digital Avatar From a Photo
Create lifelike talking faces from your photos.
VASA-1 brings unprecedented realism to virtual characters, generating synced lip movements, a full spectrum of facial expressions, and head motions. The state of the art framework produces talking faces in real-time, enhancing virtual communications across various fields from education to entertainment.
Watch the demo for eye gaze direction:
Head distance:
Emotion offsets:
Read the full report at Microsoft.com.
LeonardoAI Reveals Style Reference
Infuse your creations with the aesthetics of any reference image.
Narrative storytellers can unify visual styles. Marketers can echo visual motifs from campaigns. Designers can explore novel textures and patterns.
Watch the official demo:
Adobe’s Leap Forward in Video Enhancement
VideoGigaGAN features enhanced detail and temporal consistency in upsampled videos.
As the latest innovation in video enhancement, VideoGigaGAN addresses the critical challenges of previous video super-resolution techniques. Its sophisticated design allows for the upscaling of video content to higher resolutions without the loss of important details or introduction of flickers, promising a significant improvement in video enhancement applications.
See the dramatic results here:
See more and all the technical details on Github.
Stanford Says AI has Surpassed Human Capabilities
Annual AI Index indicates AI now performs on par with or better than humans in multiple significant cognitive areas.
From reading comprehension to visual reasoning and beyond, AI’s capabilities are expanding at an unprecedented rate. The 2023 report not only tracks these advancements but also highlights the need for new, more complex challenges to further advance the field.
Read a summary of the 500-page report at Standford.edu.
Videos of the Week
TED Talks and OpenAI’s Sora collab:
Gold Gang – C3PO & Childish Gambino by Daniel Eckler:
Gundam: A Vision of the Future by Dave Clark:
Cirque Du Freak by Ethereal Gwirl & Le Moon:
Ethereal Moon Films present "Cirque Du Freak"!@LeMoonSynth x @Ethereal_Gwirl
Coming to @fellowshiptrust @FellowshipAi soon!
Created using:
Music – @suno_ai_
SFX and Voice – @elevenlabsio
Upscales – @Magnific_AI & @topazlabs
Images – @midjourney
Animation – @runwayml &… pic.twitter.com/47Rs5yWARp— Ethereal Gwirl (@Ethereal_Gwirl) April 11, 2024
Experiments in imaginary lifeforms by Boldtron:
When odd ideas are running in circles.
Old ai experiment with #ComfyUi pic.twitter.com/Qm1t0nQKIL
— Boldtron (@edbyus) April 15, 2024
DEPICTING WATERGUNS
A new Al video experiment with #ComfyUi and pre-work of every innit image with @krea_ai realtime tool.
These are simulation tests in ai : a playful thing, controling results on sims are limited , but we'll see.
Whole video ( 1 min ) in my ig account. pic.twitter.com/J4nf7sWVOm
— Boldtron (@edbyus) April 16, 2024
1st drop with @FellowshipAi
(1000 video pieces )* MAY 23rd : Thanks @halecar2 to make this happen and Fellowship for the trust. pic.twitter.com/oiMp18Eo6A
— Boldtron (@edbyus) April 20, 2024
Some days ago I achieved this type of realism with ai. Everything done in #ComfyUI + @krea_ai realtime and scaler for some previous magic trick. You can check more creative dev at my instagram Boldtron. pic.twitter.com/Bd6HMywbca
— Boldtron (@edbyus) April 4, 2024
Join us every Wednesday for WednesdAI – a Pixel Dreams production
The images accompanying the news items in this article were generated in Midjourney using the following prompts:
