Our StyleAlign paper gets accepted to ICLR 2022 as an oral presentation. paper openreview.net/pdf?id=Qg2vi4Z… github github.com/betterze/Style…

Excited to share GANimator - our #SIGGRAPH2022 work! A single sequence is all what you need to unlock new capabilities in the character #animation domain 🤯🦀.

Excited to share our #SIGGRAPH2022 paper "CLIPasso: Semantically-Aware Object Sketching". CLIPasso converts images into sketches, with varying levels of abstraction. 🖼️🎨👩‍🎨 ✨ For more details visit our project page clipasso.github.io/clipasso/

It's been a long time coming (almost a year!!!), but I'm happy to announce that StyleGAN-NADA has been accepted to #SIGGRAPH2022 ! 🥳🥳🥳 To celebrate, we created a new Hugging Face space for you to play with: huggingface.co/spaces/rinong/… 🤗 [1/5]

This effect is driven by our eye-gaze 🤯 Google AI Come to see our #CVPR22 work - “Deep Saliency Prior for Reducing Visual Distraction”, which shows that saliency models can be used to apply various image editing effects in zero-shot setting. deep-saliency-prior.github.io

Great to see personalization introduced to large text-to-image models! Results are just stunning.

Good things come in threes! Happy to announce that MyStyle has been accepted to SIGGRAPH Asia 2022, a new and improved version is on arXiv and our source code is finally out: github.com/google/mystyle 🎉

Happy to share our latest work, “Cross-Image Attention”! 🔀🖼️🔍 We show how we can perform zero-shot appearance transfer by building on the self-attention layers of image diffusion models 😲 Great collaboration led by Yuval Alaluf and Daniel Garibi garibida.github.io/cross-image-at…

TL;DR: We use SDS from text-to-video models to animate vector-graphics sketches! Please check our project page for more details, and more penguins: livesketch.github.io

Excited to be at #ECCV2024! 🇮🇹 Presenting LazyDiffusion (lazydiffusion.github.io) — an efficient approach for local image editing with diffusion models (Tuesday, 16:30). Stop by :) DM to chat about research, Italian food recommendations, and intern opportunities at Adobe.

[1/4] Ever wondered what it would be like to use images—rather than text—to generate object and background compositions? We introduce VisualComposer, a method for compositional image generation with object-level visual prompts.

Most video models struggle to feel like real worlds. They forget what’s just out of view, slow down as videos get longer, or breaks causality. We think State Space Models are a natural fit for models with: 🧠 long-term memory across hundreds of frames ⚡ constant-speed