Summary of Compositional Image Decomposition with Diffusion Models, by Jocelin Su et al.
Compositional Image Decomposition with Diffusion Models
by Jocelin Su, Nan Liu, Yanbo Wang, Joshua B. Tenenbaum, Yilun Du
First submitted to arxiv on: 27 Jun 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: 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 Our paper presents a novel unsupervised method, Decomp Diffusion, which decomposes an image into compositional components, such as objects, lighting, shadows, and foreground. This approach infers different components in the image using diffusion models, enabling the capture of various factors like global scene descriptors or local object features. The inferred components can be flexibly composed to generate diverse scenes, even those combining elements from unseen images. Our method demonstrates promising results on natural scene decomposition and composition tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine taking a picture of a forest and breaking it down into its individual parts: trees, animals, sunlight, and shadows. Now imagine combining some of these parts with others to create a new scene that you’ve never seen before – like a bedroom full of animals under the forest’s lighting. Our team has developed a way to do just that using an unsupervised method called Decomp Diffusion. This allows us to capture different aspects of the scene and then combine them in creative ways. |
Keywords
* Artificial intelligence * Diffusion * Unsupervised
