{"ID":2833622,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.04012","arxiv_id":"2512.04012","title":"Emergent Outlier View Rejection in Visual Geometry Grounded Transformers","abstract":"Reliable 3D reconstruction from in-the-wild image collections is often hindered by \"noisy\" images-irrelevant inputs with little or no view overlap with others. While traditional Structure-from-Motion pipelines handle such cases through geometric verification and outlier rejection, feed-forward 3D reconstruction models lack these explicit mechanisms, leading to degraded performance under in-the-wild conditions. In this paper, we discover that the existing feed-forward reconstruction model, e.g., VGGT, despite lacking explicit outlier-rejection mechanisms or noise-aware training, can inherently distinguish distractor images. Through an in-depth analysis under varying proportions of synthetic distractors, we identify a specific layer that naturally exhibits outlier-suppressing behavior. Further probing reveals that this layer encodes discriminative internal representations that enable an effective noise-filtering capability, which we simply leverage to perform outlier-view rejection in feed-forward 3D reconstruction without any additional fine-tuning or supervision. Extensive experiments on both controlled and in-the-wild datasets demonstrate that this implicit filtering mechanism is consistent and generalizes well across diverse scenarios.","short_abstract":"Reliable 3D reconstruction from in-the-wild image collections is often hindered by \"noisy\" images-irrelevant inputs with little or no view overlap with others. While traditional Structure-from-Motion pipelines handle such cases through geometric verification and outlier rejection, feed-forward 3D reconstruction models...","url_abs":"https://arxiv.org/abs/2512.04012","url_pdf":"https://arxiv.org/pdf/2512.04012v1","authors":"[\"Jisang Han\",\"Sunghwan Hong\",\"Jaewoo Jung\",\"Wooseok Jang\",\"Honggyu An\",\"Qianqian Wang\",\"Seungryong Kim\",\"Chen Feng\"]","published":"2025-12-03T17:48:25Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[\"Transformer\"]","has_code":false}