DINOcular: Self-Supervised Visuospatial Representations

arXiv:2608.27226v1 Announce Type: new Abstract: We introduce a self-supervised framework for learning joint visuospatial representations from RGB-D observations. While modern vision foundation models are trained almost exclusively on RGB images, many embodied systems have access to explicit depth sensing, which provides geometric information that monocular inputs cannot recover. Our method integrates depth-derived geometric priors with a visual backbone through inter-patch and intra-patch fusion, enabling the model to encode both appearance and spatial structure efficiently. The resulting representation shows promising improvements on 3D awareness while preserving semantic transfer: it outperforms prior methods of comparable scale on multiple 3D geometry benchmarks, and remains competitive when probed for standard RGB-D semantic segmentation tasks.

Related

• Self-supervised pretraining for an iterative image size agnostic vision transformer
• RaysUp: Ultra-light Universal Feature Upsampling via Geometry-Aware Ray Representation
• Vernata: Self-Supervised Learning of LiDAR Point Representations
• Joint-Embedding Prediction of Masked Point Tubes for Self-Supervised Learning on 4D Point Cloud Videos

Source: arXiv cs.CV | 2026-08-28