Visual Lexicon: Rich Image Features in Language Space
Uloženo v:
| Vydáno v: | arXiv.org (Dec 9, 2024), p. n/a |
|---|---|
| Hlavní autor: | |
| Další autoři: | , , , |
| Vydáno: |
Cornell University Library, arXiv.org
|
| Témata: | |
| On-line přístup: | Citation/Abstract Full text outside of ProQuest |
| Tagy: |
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
MARC
| LEADER | 00000nab a2200000uu 4500 | ||
|---|---|---|---|
| 001 | 3142730777 | ||
| 003 | UK-CbPIL | ||
| 022 | |a 2331-8422 | ||
| 035 | |a 3142730777 | ||
| 045 | 0 | |b d20241209 | |
| 100 | 1 | |a Wang, XuDong | |
| 245 | 1 | |a Visual Lexicon: Rich Image Features in Language Space | |
| 260 | |b Cornell University Library, arXiv.org |c Dec 9, 2024 | ||
| 513 | |a Working Paper | ||
| 520 | 3 | |a We present Visual Lexicon, a novel visual language that encodes rich image information into the text space of vocabulary tokens while retaining intricate visual details that are often challenging to convey in natural language. Unlike traditional methods that prioritize either high-level semantics (e.g., CLIP) or pixel-level reconstruction (e.g., VAE), ViLex simultaneously captures rich semantic content and fine visual details, enabling high-quality image generation and comprehensive visual scene understanding. Through a self-supervised learning pipeline, ViLex generates tokens optimized for reconstructing input images using a frozen text-to-image (T2I) diffusion model, preserving the detailed information necessary for high-fidelity semantic-level reconstruction. As an image embedding in the language space, ViLex tokens leverage the compositionality of natural languages, allowing them to be used independently as "text tokens" or combined with natural language tokens to prompt pretrained T2I models with both visual and textual inputs, mirroring how we interact with vision-language models (VLMs). Experiments demonstrate that ViLex achieves higher fidelity in image reconstruction compared to text embeddings--even with a single ViLex token. Moreover, ViLex successfully performs various DreamBooth tasks in a zero-shot, unsupervised manner without fine-tuning T2I models. Additionally, ViLex serves as a powerful vision encoder, consistently improving vision-language model performance across 15 benchmarks relative to a strong SigLIP baseline. | |
| 653 | |a Language | ||
| 653 | |a Self-supervised learning | ||
| 653 | |a Semantics | ||
| 653 | |a Vision | ||
| 653 | |a Image quality | ||
| 653 | |a Image reconstruction | ||
| 653 | |a Natural language processing | ||
| 653 | |a Scene analysis | ||
| 653 | |a Image processing | ||
| 653 | |a Natural language | ||
| 700 | 1 | |a Zhou, Xingyi | |
| 700 | 1 | |a Fathi, Alireza | |
| 700 | 1 | |a Darrell, Trevor | |
| 700 | 1 | |a Schmid, Cordelia | |
| 773 | 0 | |t arXiv.org |g (Dec 9, 2024), p. n/a | |
| 786 | 0 | |d ProQuest |t Engineering Database | |
| 856 | 4 | 1 | |3 Citation/Abstract |u https://www.proquest.com/docview/3142730777/abstract/embedded/6A8EOT78XXH2IG52?source=fedsrch |
| 856 | 4 | 0 | |3 Full text outside of ProQuest |u http://arxiv.org/abs/2412.06774 |