{"paper":{"title":"InfoGeo: Information-Theoretic Object-Centric Learning for Cross-View Generalizable UAV Geo-Localization","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"Reformulating cross-view geo-localization as an information bottleneck that aligns object-centric structural relations across views improves robustness to domain shifts and clutter in UAV scenarios.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Hongrui Yin, Hongyang Zhang, Man On Pun, Maonnan Wang, Ziyao Wang","submitted_at":"2026-05-08T01:28:49Z","abstract_excerpt":"Cross-view geo-localization (CVGL) is fundamental for precise localization and navigation in GPS-denied environments, aiming to match ground or UAV imagery with satellite views. Existing approaches often rely on global feature alignment, but they suffer from substantial domain shifts induced by varying regional textures and weather conditions. This issue becomes even more pronounced in UAV-based scenarios, where the broader perspective inevitably introduces dense, fine-grained objects, creating significant visual clutter. To address this, we draw inspiration from Object-Centric Learning (OCL) "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"InfoGeo reformulates the optimization as an information bottleneck process with two core objectives: (i) maximizing view-invariant information by aligning the object-centric structural relations across views, and (ii) minimizing view-specific noisy signals through cross-view knowledge constraints.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That object-centric structural relations extracted from images are sufficiently view-invariant to serve as the primary signal for localization, and that cross-view knowledge constraints can be formulated to remove noise without discarding useful localization cues.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"InfoGeo reformulates cross-view geo-localization as an information bottleneck that aligns object-centric structural relations across views while minimizing view-specific noise.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Reformulating cross-view geo-localization as an information bottleneck that aligns object-centric structural relations across views improves robustness to domain shifts and clutter in UAV scenarios.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"fc2358dab5783b4dc0fb66dd1d6903a389ffb972eef29f70a4efee16b7575996"},"source":{"id":"2605.07099","kind":"arxiv","version":2},"verdict":{"id":"5d1fe868-c0b4-4f7f-ae55-d67d68a8ce95","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-11T01:24:32.515349Z","strongest_claim":"InfoGeo reformulates the optimization as an information bottleneck process with two core objectives: (i) maximizing view-invariant information by aligning the object-centric structural relations across views, and (ii) minimizing view-specific noisy signals through cross-view knowledge constraints.","one_line_summary":"InfoGeo reformulates cross-view geo-localization as an information bottleneck that aligns object-centric structural relations across views while minimizing view-specific noise.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That object-centric structural relations extracted from images are sufficiently view-invariant to serve as the primary signal for localization, and that cross-view knowledge constraints can be formulated to remove noise without discarding useful localization cues.","pith_extraction_headline":"Reformulating cross-view geo-localization as an information bottleneck that aligns object-centric structural relations across views improves robustness to domain shifts and clutter in UAV scenarios."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.07099/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T17:31:18.561374Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T12:03:22.544854Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"37f5740700ce9df4ee6de7dceefab1b4312d5bb83ad35a9e449c771fa6243eb2"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"85ca44d8b732622178a1c0fe9041d67022a417432fa366464fd13ff1f6292932"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}