Decoupling Complexity from Scale in Latent Diffusion Model
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909915421868032 |
|---|---|
| author | Zhong, Tianxiong Tian, Xingye Wang, Xuebo Jiang, Boyuan Tao, Xin Wan, Pengfei |
| author_facet | Zhong, Tianxiong Tian, Xingye Wang, Xuebo Jiang, Boyuan Tao, Xin Wan, Pengfei |
| contents | Existing latent diffusion models typically couple scale with content complexity, using more latent tokens to represent higher-resolution images or higher-frame rate videos. However, the latent capacity required to represent visual data primarily depends on content complexity, with scale serving only as an upper bound. Motivated by this observation, we propose DCS-LDM, a novel paradigm for visual generation that decouples information complexity from scale. DCS-LDM constructs a hierarchical, scale-independent latent space that models sample complexity through multi-level tokens and supports decoding to arbitrary resolutions and frame rates within a fixed latent representation. This latent space enables DCS-LDM to achieve a flexible computation-quality tradeoff. Furthermore, by decomposing structural and detailed information across levels, DCS-LDM supports a progressive coarse-to-fine generation paradigm. Experimental results show that DCS-LDM delivers performance comparable to state-of-the-art methods while offering flexible generation across diverse scales and visual qualities. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_16117 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Decoupling Complexity from Scale in Latent Diffusion Model Zhong, Tianxiong Tian, Xingye Wang, Xuebo Jiang, Boyuan Tao, Xin Wan, Pengfei Computer Vision and Pattern Recognition Existing latent diffusion models typically couple scale with content complexity, using more latent tokens to represent higher-resolution images or higher-frame rate videos. However, the latent capacity required to represent visual data primarily depends on content complexity, with scale serving only as an upper bound. Motivated by this observation, we propose DCS-LDM, a novel paradigm for visual generation that decouples information complexity from scale. DCS-LDM constructs a hierarchical, scale-independent latent space that models sample complexity through multi-level tokens and supports decoding to arbitrary resolutions and frame rates within a fixed latent representation. This latent space enables DCS-LDM to achieve a flexible computation-quality tradeoff. Furthermore, by decomposing structural and detailed information across levels, DCS-LDM supports a progressive coarse-to-fine generation paradigm. Experimental results show that DCS-LDM delivers performance comparable to state-of-the-art methods while offering flexible generation across diverse scales and visual qualities. |
| title | Decoupling Complexity from Scale in Latent Diffusion Model |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2511.16117 |