SemanticGen: Video Generation in Semantic Space
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917169311252480 |
|---|---|
| author | Bai, Jianhong Wu, Xiaoshi Wang, Xintao Fu, Xiao Zhang, Yuanxing Wang, Qinghe Shi, Xiaoyu Xia, Menghan Liu, Zuozhu Hu, Haoji Wan, Pengfei Gai, Kun |
| author_facet | Bai, Jianhong Wu, Xiaoshi Wang, Xintao Fu, Xiao Zhang, Yuanxing Wang, Qinghe Shi, Xiaoyu Xia, Menghan Liu, Zuozhu Hu, Haoji Wan, Pengfei Gai, Kun |
| contents | State-of-the-art video generative models typically learn the distribution of video latents in the VAE space and map them to pixels using a VAE decoder. While this approach can generate high-quality videos, it suffers from slow convergence and is computationally expensive when generating long videos. In this paper, we introduce SemanticGen, a novel solution to address these limitations by generating videos in the semantic space. Our main insight is that, due to the inherent redundancy in videos, the generation process should begin in a compact, high-level semantic space for global planning, followed by the addition of high-frequency details, rather than directly modeling a vast set of low-level video tokens using bi-directional attention. SemanticGen adopts a two-stage generation process. In the first stage, a diffusion model generates compact semantic video features, which define the global layout of the video. In the second stage, another diffusion model generates VAE latents conditioned on these semantic features to produce the final output. We observe that generation in the semantic space leads to faster convergence compared to the VAE latent space. Our method is also effective and computationally efficient when extended to long video generation. Extensive experiments demonstrate that SemanticGen produces high-quality videos and outperforms state-of-the-art approaches and strong baselines. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_20619 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SemanticGen: Video Generation in Semantic Space Bai, Jianhong Wu, Xiaoshi Wang, Xintao Fu, Xiao Zhang, Yuanxing Wang, Qinghe Shi, Xiaoyu Xia, Menghan Liu, Zuozhu Hu, Haoji Wan, Pengfei Gai, Kun Computer Vision and Pattern Recognition State-of-the-art video generative models typically learn the distribution of video latents in the VAE space and map them to pixels using a VAE decoder. While this approach can generate high-quality videos, it suffers from slow convergence and is computationally expensive when generating long videos. In this paper, we introduce SemanticGen, a novel solution to address these limitations by generating videos in the semantic space. Our main insight is that, due to the inherent redundancy in videos, the generation process should begin in a compact, high-level semantic space for global planning, followed by the addition of high-frequency details, rather than directly modeling a vast set of low-level video tokens using bi-directional attention. SemanticGen adopts a two-stage generation process. In the first stage, a diffusion model generates compact semantic video features, which define the global layout of the video. In the second stage, another diffusion model generates VAE latents conditioned on these semantic features to produce the final output. We observe that generation in the semantic space leads to faster convergence compared to the VAE latent space. Our method is also effective and computationally efficient when extended to long video generation. Extensive experiments demonstrate that SemanticGen produces high-quality videos and outperforms state-of-the-art approaches and strong baselines. |
| title | SemanticGen: Video Generation in Semantic Space |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.20619 |