Summary of Neural Assets: 3d-aware Multi-object Scene Synthesis with Image Diffusion Models, by Ziyi Wu et al.
Neural Assets: 3D-Aware Multi-Object Scene Synthesis with Image Diffusion Models
by Ziyi Wu, Yulia Rubanova, Rishabh Kabra, Drew A. Hudson, Igor Gilitschenski, Yusuf Aytar, Sjoerd van Steenkiste, Kelsey R. Allen, Thomas Kipf
First submitted to arxiv on: 13 Jun 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed approach in this paper introduces a novel method for controlling the 3D pose of individual objects within a scene using image diffusion models. Instead of conditioning on text tokens, the model uses per-object representations called Neural Assets to control object poses. These Neural Assets are trained to reconstruct objects from reference images and condition on object poses from target frames. This allows for learning disentangled appearance and pose features. The approach combines visual and 3D pose representations in a sequence-of-tokens format, allowing for fine-grained 3D pose and placement control of individual objects. The model achieves state-of-the-art results on both synthetic and real-world video datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about finding ways to control the position and movement of objects within images or videos using special computer models called image diffusion models. Instead of using words, this method uses special pictures that describe what each object looks like. These pictures help the model understand how to move the objects around in new scenes. This allows for more detailed control over where objects are placed and how they move. The results show that this approach works well on both fake and real videos. |
Keywords
» Artificial intelligence » Diffusion
