AI Art Weekly #104
Hello there, my fellow dreamers, and welcome to issue #104 of AI Art Weekly! π
Models are getting better and better and moats are vanishing as fast as they are created. At the beginning of the week I stumbled upon a tool that lets you quickly add text behind subjects in images, so I was curious if I could replicate it with open-source models that run 100% in the browser. And I did: Text Behind Image. Itβs free and even runs on mobile devices, so give it a try and let me know what you think!
Also added another 10 high-quality SREF codes to PROMPTCACHE this week. If youβre looking for inspiration, you should definitely check them out.
Tired of no-code papers? Bookmark them directly in the newsletter below and get notified once their code is released. Unlock now.
Cover Challenge π¨
For the next cover Iβm looking for threshold inspired submissions! Reward is a rare role in our Discord community which lets you vote in the finals. Rulebook can be found here and images can be submitted here.
News & Papers
Highlights
RMBG-2.0
BRIA released the 2nd version of their background removal model RMBG-2.0. Itβs a new state of the art model that can remove backgrounds from images. My Text Behind Image tool currently runs the predecessor model RMBG-1.4. Not sure if 2.0 also works on mobile devices, but Iβll give it a try once I figure out how to implement it.
RMBG-2.0 examples
3D
BPT: Scaling Mesh Generation via Compressive Tokenization
Scaling Mesh Generation via Compressive Tokenization can generate high-quality meshes with over 8,000 faces.
Scaling Mesh Generation via Compressive Tokenization example
GaussianAnything: Interactive Point Cloud Latent Diffusion for 3D Generation
GaussianAnything can generate high-quality 3D objects from single images or text prompts. It uses a Variational Autoencoder and a cascaded latent diffusion model for effective 3D editing.
GaussianAnything example
Edify 3D: Scalable High-Quality 3D Asset Generation
Edify 3D can generate high-quality 3D assets from text descriptions. It uses a diffusion model to create detailed quad-mesh topologies and high-resolution textures in under 2 minutes.
Edify 3D example
StdGEN: Semantic-Decomposed 3D Character Generation from Single Images
StdGEN can generate high-quality 3D characters from a single image in just three minutes. It breaks down characters into parts like body, clothes, and hair, using a transformer-based model for great results in 3D anime character generation.
StdGEN example
ProEdit: Simple Progression is All You Need for High-Quality 3D Scene Editing
ProEdit can edit 3D scenes by breaking tasks into smaller steps, which helps reduce errors in different views. It has a smart scheduler that considers task difficulty and lets users control how aggressive the edits are, achieving high-quality results without extra tools.
ProEdit example
DanceFusion: A Spatio-Temporal Skeleton Diffusion Transformer for Audio-Driven Dance Motion Reconstruction
DanceFusion can generate and reconstruct dance pose movements that match music.
DanceFusion example
Image
MagicQuill: An Intelligent Interactive Image Editing System
MagicQuill enables efficient image editing with a simple interface that lets users easily insert elements and change colors. It uses a large language model to understand editing intentions in real time, improving the quality of the results.
MagicQuill example
Add-it: Training-Free Object Insertion in Images With Pretrained Diffusion Models
Add-it can add objects to images based on text prompts without extra training. It uses a smart attention system for natural placement and consistency, achieving top results in image insertion tasks.
Add-it example
SeedEdit: Align Image Re-Generation to Image Editing
SeedEdit is a image model that can revise images based on text prompts while keeping a balance between changing and maintaining the original image. It allows for high-resolution editing and supports various changes like local replacements, geometric transformations, and style adjustments.
SeedEdit example
Video
DimensionX: Create Any 3D and 4D Scenes from a Single Image with Controllable Video Diffusion
DimensionX can generate photorealistic 3D and 4D scenes from a single image using controllable video diffusion.
DimensionX example
SG-I2V: Self-Guided Trajectory Control in Image-to-Video Generation
SG-I2V can control object and camera motion in image-to-video generation using bounding boxes and trajectories
SG-I2V example
MVideo: Motion Control for Enhanced Complex Action Video Generation
MVideo can generate long videos with smooth actions by using mask sequences for motion control.
Motion Control for Enhanced Complex Action Video Generation example
ReCapture: Generative Video Camera Controls for User-Provided Videos using Masked Video Fine-Tuning
ReCapture can generate new videos with different camera angles from a single video.
ReCapture example
Also interesting
--sref 1464347777. Pastel Phantasm is a dreamy and vintage-inspired style that combines ethereal elements and soft pastel colors to create atmospheric, photorealistic images that evoke contemplation and a sense of nostalgia.
@nutlope built an AI tool that lets you create professional logos in seconds using Flux. 100% free and open source.
βGN dreamers β¨β by me.
And that my fellow dreamers, concludes yet another AI Art weekly issue. Please consider supporting this newsletter by:
- Sharing it πβ€οΈ
- Following me on Twitter: @dreamingtulpa
- Buying me a coffee (I could seriously use it, putting these issues together takes me 8-12 hours every Friday π )
- Buying my Midjourney prompt collection on PROMPTCACHE π
- Buying access to AI Art Weekly Premium π
Reply to this email if you have any feedback or ideas for this newsletter.
Thanks for reading and talk to you next week!
β dreamingtulpa