Object Concepts Emerge from Motion

• URL: http://arxiv.org/abs/2609.04348v1
• Date: Thu, 03 Sep 2026 18:11:53 GMT
• Title: Object Concepts Emerge from Motion
• Abstract summary: We present a biologically inspired framework that learns object-centric representations for single images from raw videos.<n>We first obtain 195 million pseudo-labeled frames from 7,163 hours of driving and web videos, then expand the supervision to 421 million frames.<n>We train encoders up to Swin-H and distill the learned representations into a family of Swin backbones.
• Score: 25.362771842390874
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Object-centric visual representations are important for physical-world perception, but existing visual pretraining methods often capture semantic categories without preserving the identity and coherence of individual instances. We present a biologically inspired framework that learns object-centric representations for single images from raw videos. Our approach uses motion boundaries as a source of object-level grouping: off-the-shelf optical flow and clustering produce pseudo-instance masks, which supervise a single-image encoder with pixel-level pairwise metric learning. The framework requires neither human annotations nor camera calibration. We first obtain 195 million pseudo-labeled frames from 7,163 hours of driving and web videos, then expand the supervision to 421 million frames with Motion-Verified Self-Training, which combines model proposals with motion evidence. We train encoders up to Swin-H and distill the learned representations into a family of Swin backbones. Across monocular depth estimation, 3D object detection, 3D occupancy prediction, and end-to-end planning, the resulting models achieve competitive or superior performance relative to supervised and self-supervised pretraining baselines, with particularly strong transfer on geometry- and instance-sensitive tasks. These results show that motion-derived supervision can teach static image encoders to represent visual instances, providing a complementary direction for scalable visual pretraining.

Related papers

• Vision Pretraining for Dense Spatial Perception [69.54043654361772]
  We study vision pretraining through a boundary-centric lens, motivated by the premise that boundaries and shape discontinuities offer essential cues for perceiving geometric properties.<n>Our findings reveal that boundary modeling goes beyond simple line segments and instead serves as a scalable pretraining principle for learning spatially structured visual representations.
  arXiv  Detail & Related papers  (2026-07-06T15:56:14Z)
• PAOLI: Pose-free Articulated Object Learning from Sparse-view Images [27.16160315662701]
  We present a novel framework for learning articulated object representations from sparse-view, unposed images.<n>Our approach operates with as few as four views per articulation and no camera supervision.
  arXiv  Detail & Related papers  (2025-09-04T14:51:03Z)
• Object Concepts Emerge from Motion [24.73461163778215]
  We propose a biologically inspired framework for learning object-centric visual representations in an unsupervised manner.<n>Our key insight is that motion boundary serves as a strong signal for object-level grouping.<n>Our framework is fully label-free and does not rely on camera calibration, making it scalable to large-scale unstructured video data.
  arXiv  Detail & Related papers  (2025-05-27T18:09:02Z)
• Perception Encoder: The best visual embeddings are not at the output of the network [70.86738083862099]
  We introduce Perception (PE), a vision encoder for image and video understanding trained via simple vision-language learning.<n>We find that contrastive vision-language training alone can produce strong, general embeddings for all of these downstream tasks.<n>Together, our PE family of models achieves best-in-class results on a wide variety of tasks.
  arXiv  Detail & Related papers  (2025-04-17T17:59:57Z)
• Heuristic Vision Pre-Training with Self-Supervised and Supervised Multi-Task Learning [0.0]
  We propose a novel pre-training framework by adopting both self-supervised and supervised visual pre-text tasks in a multi-task manner. Results show that our pre-trained models can deliver results on par with or better than state-of-the-art (SOTA) results on multiple visual tasks.
  arXiv  Detail & Related papers  (2023-10-11T14:06:04Z)
• AutoDecoding Latent 3D Diffusion Models [95.7279510847827]
  We present a novel approach to the generation of static and articulated 3D assets that has a 3D autodecoder at its core. The 3D autodecoder framework embeds properties learned from the target dataset in the latent space. We then identify the appropriate intermediate volumetric latent space, and introduce robust normalization and de-normalization operations.
  arXiv  Detail & Related papers  (2023-07-07T17:59:14Z)
• Attentive and Contrastive Learning for Joint Depth and Motion Field Estimation [76.58256020932312]
  Estimating the motion of the camera together with the 3D structure of the scene from a monocular vision system is a complex task. We present a self-supervised learning framework for 3D object motion field estimation from monocular videos.
  arXiv  Detail & Related papers  (2021-10-13T16:45:01Z)
• Learning Monocular Depth in Dynamic Scenes via Instance-Aware Projection Consistency [114.02182755620784]
  We present an end-to-end joint training framework that explicitly models 6-DoF motion of multiple dynamic objects, ego-motion and depth in a monocular camera setup without supervision. Our framework is shown to outperform the state-of-the-art depth and motion estimation methods.
  arXiv  Detail & Related papers  (2021-02-04T14:26:42Z)
• Self-Supervised Representation Learning from Flow Equivariance [97.13056332559526]
  We present a new self-supervised learning representation framework that can be directly deployed on a video stream of complex scenes. Our representations, learned from high-resolution raw video, can be readily used for downstream tasks on static images.
  arXiv  Detail & Related papers  (2021-01-16T23:44:09Z)
• Lift, Splat, Shoot: Encoding Images From Arbitrary Camera Rigs by Implicitly Unprojecting to 3D [100.93808824091258]
  We propose a new end-to-end architecture that directly extracts a bird's-eye-view representation of a scene given image data from an arbitrary number of cameras. Our approach is to "lift" each image individually into a frustum of features for each camera, then "splat" all frustums into a bird's-eye-view grid. We show that the representations inferred by our model enable interpretable end-to-end motion planning by "shooting" template trajectories into a bird's-eye-view cost map output by our network.
  arXiv  Detail & Related papers  (2020-08-13T06:29:01Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.