{"paper":{"title":"WildPose: A Unified Framework for Robust Pose Estimation in the Wild","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"WildPose unifies monocular pose estimation to stay accurate in dynamic scenes without losing ground on static or low-motion ones.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Iro Armeni, Jianhao Zheng, Liyuan Zhu, Zihan Zhu","submitted_at":"2026-05-12T21:39:44Z","abstract_excerpt":"Estimating camera pose in dynamic environments is a critical challenge, as most visual SLAM and SfM methods assume static scenes. While recent dynamic-aware methods exist, they are often not unified: semantic-based approaches are brittle, per-sequence optimization methods fail on short sequences, and other learned models may degrade on static-only scenes. We present WildPose, a unified monocular pose estimation framework that is robust in dynamic environments while maintaining state-of-the-art performance on static and low-ego-motion datasets. Our key insight is to connect two powerful paradig"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"WildPose consistently outperforms prior methods across dynamic (Wild-SLAM, Bonn), static (TUM, 7-Scenes), and low-ego-motion (Sintel) benchmarks.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That a frozen pre-trained MASt3R backbone supplies sufficiently rich and general 3D-aware features for both the update operator and motion mask detector to deliver unified robustness without scene-specific retraining or failure modes in unseen dynamics.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"WildPose unifies feedforward 3D features from MASt3R with differentiable bundle adjustment for robust monocular pose estimation across dynamic, static, and low-ego-motion scenes.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"WildPose unifies monocular pose estimation to stay accurate in dynamic scenes without losing ground on static or low-motion ones.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"d683c8c450de5c50dc77464a74bebc415c1282a52bcd2a56b2bb5de528fb6e9b"},"source":{"id":"2605.12774","kind":"arxiv","version":1},"verdict":{"id":"3e7d3a39-99a7-4144-96f7-1deb3704d861","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T20:34:19.048067Z","strongest_claim":"WildPose consistently outperforms prior methods across dynamic (Wild-SLAM, Bonn), static (TUM, 7-Scenes), and low-ego-motion (Sintel) benchmarks.","one_line_summary":"WildPose unifies feedforward 3D features from MASt3R with differentiable bundle adjustment for robust monocular pose estimation across dynamic, static, and low-ego-motion scenes.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That a frozen pre-trained MASt3R backbone supplies sufficiently rich and general 3D-aware features for both the update operator and motion mask detector to deliver unified robustness without scene-specific retraining or failure modes in unseen dynamics.","pith_extraction_headline":"WildPose unifies monocular pose estimation to stay accurate in dynamic scenes without losing ground on static or low-motion ones."},"references":{"count":72,"sample":[{"doi":"","year":2018,"title":"Fácil, Javier Civera, and José Neira","work_id":"490fd45e-a87c-4270-bc0b-b7b35baa0520","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2012,"title":"A naturalistic open source movie for optical flow evaluation","work_id":"d4c5c49b-f5f0-41bc-8b80-df8b660bc546","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Orb-slam3: An accu- rate open-source library for visual, visual–inertial, and mul- timap slam.IEEE transactions on robotics, 37(6):1874–1890,","work_id":"27d28ed6-c2e5-4fa0-bc3f-0edd27a4f2c0","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Video depth anything: Consistent depth estimation for super-long videos","work_id":"79a25c43-dc22-412e-8807-555f77f0a128","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Back on track: Bundle adjustment for dynamic scene reconstruc- tion","work_id":"86b6761f-a7bf-4610-b5c8-68391da12516","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":72,"snapshot_sha256":"d0b4eb8eaeba7947de11331cc6a63b4b34dfbaf223ed5f056d5ca068b7b83726","internal_anchors":6},"formal_canon":{"evidence_count":2,"snapshot_sha256":"64034ed3ff1e98c109ce95fc923ca54fe3280b5f6d2f79e9ad5a3a7bfa97283d"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}