Abstract
Self-supervised image encoders such as DINO have recently gained significant interest for learning robust visual features without labels. However, most SSL methods train on static images and miss the temporal cues inherent in videos. We introduce a video-distilled single-image encoder trained to predict the next-frame representation from the current frame. This simple objective injects 3D spatial and temporal priors without optical flow or tracking. When pre-training on a single 2-hour video, our approach raises the mean Intersection-over-Union (mIoU) on ADE20K from 35.0 (DoRA) to 36.4 while remaining a drop-in replacement for image-only pipelines. Our results highlight video self-distillation as a lightweight route to geometry-aware perception an essential ingredient for physically plausible world models and Physical AI.
Abstract (translated)
近期，像DINO这样的自监督图像编码器因能够在无标签的情况下学习稳健的视觉特征而备受关注。然而，大多数自监督学习（SSL）方法都是在静态图像上进行训练，忽视了视频中固有的时间线索。为此，我们提出了一种基于单帧预测下一帧表示的视频蒸馏单一图像编码器。这种简单的目标无需光学流或跟踪技术就能注入3D空间和时间先验信息。当仅使用一段2小时长的视频进行预训练时，我们的方法使ADE20K数据集上的平均交并比（mIoU）从35.0（DoRA结果）提升到了36.4，并且可以作为图像管道中的直接替代方案。我们的成果强调了视频自蒸馏作为一种轻量级的方法，有助于几何感知的视觉理解，这对于构建物理上合理的世界模型和实现物理人工智能至关重要。
URL
https://arxiv.org/abs/2507.19272
