LSSGen: Leveraging Latent Space Scaling in Flow and Diffusion for Efficient Text to Image Generation
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866912496432971776 |
|---|---|
| author | Tang, Jyun-Ze Hsu, Chih-Fan Li, Jeng-Lin Chang, Ming-Ching Chen, Wei-Chao |
| author_facet | Tang, Jyun-Ze Hsu, Chih-Fan Li, Jeng-Lin Chang, Ming-Ching Chen, Wei-Chao |
| contents | Flow matching and diffusion models have shown impressive results in text-to-image generation, producing photorealistic images through an iterative denoising process. A common strategy to speed up synthesis is to perform early denoising at lower resolutions. However, traditional methods that downscale and upscale in pixel space often introduce artifacts and distortions. These issues arise when the upscaled images are re-encoded into the latent space, leading to degraded final image quality. To address this, we propose {\bf Latent Space Scaling Generation (LSSGen)}, a framework that performs resolution scaling directly in the latent space using a lightweight latent upsampler. Without altering the Transformer or U-Net architecture, LSSGen improves both efficiency and visual quality while supporting flexible multi-resolution generation. Our comprehensive evaluation covering text-image alignment and perceptual quality shows that LSSGen significantly outperforms conventional scaling approaches. When generating $1024^2$ images at similar speeds, it achieves up to 246\% TOPIQ score improvement. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_16154 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | LSSGen: Leveraging Latent Space Scaling in Flow and Diffusion for Efficient Text to Image Generation Tang, Jyun-Ze Hsu, Chih-Fan Li, Jeng-Lin Chang, Ming-Ching Chen, Wei-Chao Computer Vision and Pattern Recognition Artificial Intelligence Flow matching and diffusion models have shown impressive results in text-to-image generation, producing photorealistic images through an iterative denoising process. A common strategy to speed up synthesis is to perform early denoising at lower resolutions. However, traditional methods that downscale and upscale in pixel space often introduce artifacts and distortions. These issues arise when the upscaled images are re-encoded into the latent space, leading to degraded final image quality. To address this, we propose {\bf Latent Space Scaling Generation (LSSGen)}, a framework that performs resolution scaling directly in the latent space using a lightweight latent upsampler. Without altering the Transformer or U-Net architecture, LSSGen improves both efficiency and visual quality while supporting flexible multi-resolution generation. Our comprehensive evaluation covering text-image alignment and perceptual quality shows that LSSGen significantly outperforms conventional scaling approaches. When generating $1024^2$ images at similar speeds, it achieves up to 246\% TOPIQ score improvement. |
| title | LSSGen: Leveraging Latent Space Scaling in Flow and Diffusion for Efficient Text to Image Generation |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2507.16154 |