Video Generation Models are General-Purpose Vision Learners
Quick Answer
This paper shows that GenCeption, a video generation model, demonstrates state-of-the-art performance across various vision tasks, outperforming specialized models like DepthAnything3 and D4RT with significantly less training data.
Quick Take
GenCeption, a video generation model, demonstrates state-of-the-art performance across various vision tasks, outperforming specialized models like DepthAnything3 and D4RT with significantly less training data. This model leverages a pre-trained video generative diffusion backbone, showcasing exceptional data efficiency and emergent behaviors in real-world applications.
Key Points
- GenCeption achieves state-of-the-art results in depth and camera pose estimation.
- It requires 7 to 500 times less training data than leading models like D4RT.
- The model generalizes well to real-world footage from synthetic human videos.
- Video generation serves as a foundational approach for general visual intelligence.
- GenCeption outperforms alternative pretraining methods like V-JEPA and Video MAE.
Paper Resources
📖 Reader Mode
~2 min readAuthors:Letian Wang, Chuhan Zhang, Rishabh Kabra, Jasper Uijlings, Steven Waslander, Andrew Zisserman, Joao Carreira, Kaiming He, Misha Andriluka, Eduard Gabriel Bazavan, Andrei Zanfir, Cristian Sminchisescu
Abstract:Driven by next-token prediction, NLP shifted from task-specific models into powerful generalist foundation models. What, then, is the equivalent catalyst needed to achieve a general-purpose model in computer vision? In this paper, we contend that large-scale text-to-video generation serves as a strong pre-training paradigm for computer vision, providing the necessary spatiotemporal priors, vision-language alignment, and scalability required for general visual intelligence. We introduce GenCeption, which leverages a pre-trained video generative diffusion backbone to define a feed-forward perception model, capable of performing various vision tasks steered by text instructions. Empirical results demonstrate that GenCeption achieves state-of-the-art performance across a diverse suite of tasks, including depth, surface normal, and camera pose estimation, expression-referring segmentation, and 3D keypoint prediction, often matching or surpassing specialized models (e.g. DepthAnything3, SAM3, D4RT, VGGT-Omega, Sapiens, David, Genmo, and Lotus-2). Furthermore, the video generative pretrained backbone outperforms alternative pretraining paradigms (e.g., V-JEPA, and Video MAE) under comparable settings. Importantly, GenCeption exhibits preliminary data and model scaling properties along with exceptional data efficiency, where it achieves comparable performance with leading models like D4RT and VGGT-Omega with 7 to 500 less training data. Finally, GenCeption also exhibits intriguing emergent behaviors: a model trained exclusively on synthetic human videos generalizes to real-world footage and out-of-distribution object categories (e.g., animals and robots). These findings suggest that video generation is not merely a synthesis tool, but a foundational path toward generalist vision intelligence for the physical world. Project page: this https URL
| Comments: | ECCV 2026 |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2607.09024 [cs.CV] |
| (or arXiv:2607.09024v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2607.09024 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Letian Wang [view email]
[v1]
Fri, 10 Jul 2026 01:09:06 UTC (17,025 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →ProMoE-FL: Prototype-conditioned Mixture of Experts for Multimodal Federated Learning with Missing Modalities
ProMoE-FL introduces a Prototype-conditioned Mixture-of-Experts framework for multimodal federated learning, effectively addressing missing modalities. It outperforms existing methods on four chest X-ray datasets, demonstrating superior feature synthesis capabilities in both homogeneous and heterogeneous settings.