{"paper":{"title":"PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow","license":"http://creativecommons.org/licenses/by/4.0/","headline":"PVRF uses zero-shot weather perceptions from frozen vision-language models to guide a velocity-constrained rectified flow that refines restoration anchors for multiple adverse degradations.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Guangtao Zhai, Guanhua Zhao, Han Zhou, Jun Chen, Shahab Asoodeh, Terry Ji, Wei Dong, Xiaohong Liu, Yulun Zhang","submitted_at":"2026-05-13T19:07:14Z","abstract_excerpt":"Adverse weather removal (AWR) in real-world images remains challenging due to heterogeneous and unseen degradations, while distortion-driven training often yields overly smooth results. We propose PVRF, a unified framework that integrates zero-shot soft weather perceptions with velocity-constrained rectified-flow refinement. PVRF introduces an AWR-specific question answering module (AWR-QA) that uses frozen vision--language models (VLMs) to estimate soft probabilities of weather types and low-level attribute scores. These perceptions condition restoration networks via attribute-modulated norma"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"PVRF improves both fidelity and perceptual quality over state-of-the-art baselines, with strong cross-dataset generalization on single and combined degradations.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"Zero-shot soft weather perceptions produced by frozen VLMs via the AWR-QA module are sufficiently accurate and informative to condition the restoration networks effectively through AMN and WWA.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"PVRF combines zero-shot VLM-based weather perception with perception-adaptive rectified flow refinement to achieve all-in-one adverse weather removal with improved fidelity and cross-dataset generalization.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"PVRF uses zero-shot weather perceptions from frozen vision-language models to guide a velocity-constrained rectified flow that refines restoration anchors for multiple adverse degradations.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"9b776bbb19dd19ca86839ad690679e00b9dec62a88a17ee530cbe98526d7a103"},"source":{"id":"2605.14045","kind":"arxiv","version":1},"verdict":{"id":"6cc8dc28-3b0d-4924-86f6-2a7b1e8b2e6b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T05:41:19.488862Z","strongest_claim":"PVRF improves both fidelity and perceptual quality over state-of-the-art baselines, with strong cross-dataset generalization on single and combined degradations.","one_line_summary":"PVRF combines zero-shot VLM-based weather perception with perception-adaptive rectified flow refinement to achieve all-in-one adverse weather removal with improved fidelity and cross-dataset generalization.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"Zero-shot soft weather perceptions produced by frozen VLMs via the AWR-QA module are sufficiently accurate and informative to condition the restoration networks effectively through AMN and WWA.","pith_extraction_headline":"PVRF uses zero-shot weather perceptions from frozen vision-language models to guide a velocity-constrained rectified flow that refines restoration anchors for multiple adverse degradations."},"references":{"count":49,"sample":[{"doi":"","year":2018,"title":"The perception-distortion tradeoff","work_id":"8f7c1692-5452-4dc1-97b6-56c11153d5f8","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Unirestore: Unified perceptual and task-oriented image restoration model using diffusion prior","work_id":"acbdda1c-5bf7-432a-bff7-a505cc3f3e73","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Simple baselines for image restoration","work_id":"0e714e9d-2b6b-4cda-ae2e-3c9d63bb1d5d","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"All snow removed: Single image desnowing algorithm using hierarchical dual-tree complex wavelet representation and contradict channel loss","work_id":"ff5d9456-46bf-4231-839e-d2b356bf1c0a","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Bio-inspired image restoration","work_id":"f2b9bd5f-9bf0-42f2-a338-e751fc8c71e5","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":49,"snapshot_sha256":"e33c2f6542eb037ff4286df1eeb0f68be18962cfbe54fdb3af14c5733bc49bae","internal_anchors":1},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}