Summary of Promptdresser: Improving the Quality and Controllability Of Virtual Try-on Via Generative Textual Prompt and Prompt-aware Mask, by Jeongho Kim et al.
PromptDresser: Improving the Quality and Controllability of Virtual Try-On via Generative Textual Prompt and Prompt-aware Mask
by Jeongho Kim, Hoiyeong Jin, Sunghyun Park, Jaegul Choo
First submitted to arxiv on: 22 Dec 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
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 PromptDresser model leverages large multimodal models (LMMs) to enable high-quality and versatile manipulation of person-clothing images based on generative text prompts. The model tackles the text-editable virtual try-on task, which involves changing the clothing item and editing wearing styles according to text descriptions. To address this challenge, PromptDresser designs rich text descriptions for paired person-clothing data, addresses conflicts between textual information and generation, and adapts inpainting masks aligned with text prompts. This approach utilizes LMMs via in-context learning to generate detailed text descriptions and pose details using minimal human cost. The model is evaluated on its ability to enhance image quality and convey clothing details that are difficult to capture through images alone. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper proposes a new way to change clothes and styles in virtual try-on pictures based on text prompts. It uses a special kind of artificial intelligence (AI) called a large multimodal model, which helps the AI learn from lots of data. The AI is trained to understand what people are wearing and what they want to wear, and it can even add details like pose and style to the picture. This makes the virtual try-on pictures more realistic and fun to use. |
