English

Motion-Refined DINOSAUR for Unsupervised Multi-Object Discovery

Computer Vision and Pattern Recognition 2025-09-03 v1

Abstract

Unsupervised multi-object discovery (MOD) aims to detect and localize distinct object instances in visual scenes without any form of human supervision. Recent approaches leverage object-centric learning (OCL) and motion cues from video to identify individual objects. However, these approaches use supervision to generate pseudo labels to train the OCL model. We address this limitation with MR-DINOSAUR -- Motion-Refined DINOSAUR -- a minimalistic unsupervised approach that extends the self-supervised pre-trained OCL model, DINOSAUR, to the task of unsupervised multi-object discovery. We generate high-quality unsupervised pseudo labels by retrieving video frames without camera motion for which we perform motion segmentation of unsupervised optical flow. We refine DINOSAUR's slot representations using these pseudo labels and train a slot deactivation module to assign slots to foreground and background. Despite its conceptual simplicity, MR-DINOSAUR achieves strong multi-object discovery results on the TRI-PD and KITTI datasets, outperforming the previous state of the art despite being fully unsupervised.

Keywords

Cite

@article{arxiv.2509.02545,
  title  = {Motion-Refined DINOSAUR for Unsupervised Multi-Object Discovery},
  author = {Xinrui Gong and Oliver Hahn and Christoph Reich and Krishnakant Singh and Simone Schaub-Meyer and Daniel Cremers and Stefan Roth},
  journal= {arXiv preprint arXiv:2509.02545},
  year   = {2025}
}

Comments

To appear at ICCVW 2025. Xinrui Gong and Oliver Hahn - both authors contributed equally. Code: https://github.com/visinf/mr-dinosaur

R2 v1 2026-07-01T05:17:45.878Z