{"ID":2855896,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.12764","arxiv_id":"2510.12764","title":"AnyUp: Universal Feature Upsampling","abstract":"We introduce AnyUp, a method for feature upsampling that can be applied to any vision feature at any resolution, without encoder-specific training. Existing learning-based upsamplers for features like DINO or CLIP need to be re-trained for every feature extractor and thus do not generalize to different feature types at inference time. In this work, we propose an inference-time feature-agnostic upsampling architecture to alleviate this limitation and improve upsampling quality. In our experiments, AnyUp sets a new state of the art for upsampled features, generalizes to different feature types, and preserves feature semantics while being efficient and easy to apply to a wide range of downstream tasks.","short_abstract":"We introduce AnyUp, a method for feature upsampling that can be applied to any vision feature at any resolution, without encoder-specific training. Existing learning-based upsamplers for features like DINO or CLIP need to be re-trained for every feature extractor and thus do not generalize to different feature types at...","url_abs":"https://arxiv.org/abs/2510.12764","url_pdf":"https://arxiv.org/pdf/2510.12764v2","authors":"[\"Thomas Wimmer\",\"Prune Truong\",\"Marie-Julie Rakotosaona\",\"Michael Oechsle\",\"Federico Tombari\",\"Bernt Schiele\",\"Jan Eric Lenssen\"]","published":"2025-10-14T17:45:17Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.LG\"]","methods":"[]","has_code":false}