Summary of Omniprism: Learning Disentangled Visual Concept For Image Generation, by Yangyang Li et al.
OmniPrism: Learning Disentangled Visual Concept for Image Generation
by Yangyang Li, Daqing Liu, Wu Liu, Allen He, Xinchen Liu, Yongdong Zhang, Guoqing Jin
First submitted to arxiv on: 16 Dec 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 OmniPrism approach aims to improve creative visual concept generation by disentangling concept representations from reference images. The method learns disentangled concepts using natural language guidance and trains a diffusion model to incorporate these concepts. A paired concept disentangled dataset (PCD-200K) is constructed to disentangle concepts with different semantics. The COD training pipeline is used to learn disentangled concept representations, which are then injected into additional diffusion cross-attention layers for generation. Experimental results show that OmniPrism can generate high-quality, concept-disentangled results with high fidelity to text prompts and desired concepts. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary OmniPrism is a new way to create images based on ideas from reference pictures. Right now, most image creation methods have limitations when dealing with multiple ideas or concepts in one picture. To solve this problem, researchers developed OmniPrism, which learns to separate different concepts and then uses these separated concepts to generate new images. |
Keywords
» Artificial intelligence » Cross attention » Diffusion » Diffusion model » Semantics
