Papers › Toward unsupervised, multi-object discovery in large-scale image collections

Toward unsupervised, multi-object discovery in large-scale image collections

6 Jul 2020ECCV 2020 8arXiv:2007.02662archive 2025-07-28

Huy V. Vo, Patrick Pérez, Jean Ponce

This paper addresses the problem of discovering the objects present in a collection of images without any supervision. We build on the optimization approach of Vo et al. (CVPR'19) with several key novelties: (1) We propose a novel saliency-based region proposal algorithm that achieves significantly higher overlap with ground-truth objects than other competitive methods. This procedure leverages off-the-shelf CNN features trained on classification tasks without any bounding box information, but is otherwise unsupervised. (2) We exploit the inherent hierarchical structure of proposals as an effective regularizer for the approach to object discovery of Vo et al., boosting its performance to significantly improve over the state of the art on several standard benchmarks. (3) We adopt a two-stage strategy to select promising proposals using small random sets of images before using the whole image collection to discover the objects it depicts, allowing us to tackle, for the first time (to the best of our knowledge), the discovery of multiple objects in each one of the pictures making up datasets with up to 20,000 images, an over five-fold increase compared to existing methods, and a first step toward true large-scale unsupervised image interpretation.

PaperPDFConference PDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

huyvvo/rOSD officialGPL-3.0 report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Multi-object colocalizationMulti-object discoveryObject DiscoveryRegion ProposalSingle-object colocalizationSingle-object discovery

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Multi-object colocalization VOC12 rOSD Detection Rate 51.5 #1 of 1 Archive leaderboard report
Multi-object colocalization VOC_all rOSD Detection Rate 49.4 #1 of 1 Archive leaderboard report
Multi-object discovery COCO_20k Large-scale rOSD Detection Rate 12.0 #1 of 1 Archive leaderboard report
Multi-object discovery VOC12 Large-scale rOSD Detection Rate 41.2 #1 of 2 Archive leaderboard report
Multi-object discovery VOC12 rOSD Detection Rate 40.4 #2 of 2 Archive leaderboard report
Multi-object discovery VOC_all Large-scale rOSD Detection Rate 38.3 #1 of 2 Archive leaderboard report
Multi-object discovery VOC_all rOSD Detection Rate 37.6 #2 of 2 Archive leaderboard report
Single-object discovery COCO_20k rOSD + CAD CorLoc 53.0 #7 of 10 Archive leaderboard report
Single-object discovery COCO_20k rOSD CorLoc 48.5 #9 of 10 Archive leaderboard report
Single-object discovery Object Discovery rOSD CorLoc 89.2 #1 of 2 Archive leaderboard report
Single-object discovery VOC12 Large-scale rOSD CorLoc 51.9 #1 of 2 Archive leaderboard report
Single-object discovery VOC12 rOSD CorLoc 51.2 #2 of 2 Archive leaderboard report
Single-object discovery VOC_6x2 rOSD CorLoc 72.5 #1 of 2 Archive leaderboard report
Single-object discovery VOC_all Large-scale rOSD CorLoc 49.4 #1 of 3 Archive leaderboard report
Single-object discovery VOC_all rOSD CorLoc 49.3 #2 of 3 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections