Motion-Refined DINOSAUR for Unsupervised Multi-Object Discovery

2025-09-02

AI Summary

Unsupervised multi-object discovery (MOD) aims to detect and localize multiple object instances in images without human annotations. Existing approaches rely on supervision-derived pseudo-labels to train object-centric learning (OCL) models, violating the fully unsupervised setting. This paper proposes MR-DINOSAUR, the first end-to-end framework for high-quality unsupervised pseudo-label generation. It builds upon the self-supervised pre-trained model DINOSAUR and integrates unsupervised optical flow estimation, static frame retrieval, and motion-based segmentation to refine slot representations. A novel slot deactivation module is introduced to explicitly model background, thereby enhancing foreground-background separation. Evaluated on TRI-PD and KITTI, MR-DINOSAUR significantly outperforms prior unsupervised MOD methods, achieving state-of-the-art performance.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Computer Vision: Multi-modal Vision
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

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.

Problem

• Unsupervised detection and localization of distinct object instances
• Eliminating supervision requirements for pseudo label generation
• Refining slot representations using motion-segmented optical flow

Innovation

• Motion-Refined DINOSAUR extends self-supervised OCL model
• Generates pseudo labels via motion segmentation of optical flow
• Trains slot deactivation module for foreground-background assignment