{"paper":{"title":"PRISM: Rethinking Atmospheric Scattering Reconstruction as a Unified Understanding and Restoration Model for Real-world Dehazing","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"PRISM jointly reconstructs clear scenes and scattering variables to improve real-world image dehazing.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chengyu Fang, Chenyang Zhu, Chubin Chen, Chunming He, Hongqiu Wang, Longxiang Tang, Sina Farsiu, Xiu Li, Yuelin Zhang","submitted_at":"2026-04-08T13:01:30Z","abstract_excerpt":"Real-world image dehazing (RID) aims to remove haze-induced degradation from real scenes. This task remains challenging due to non-uniform haze distribution, spatially varying color shifts, and the scarcity of paired real hazy-clean data. In PRISM, we propose Proximal Scattering Atmosphere Reconstruction (PSAR), a physically structured framework that jointly reconstructs the clear scene and scattering variables under the atmospheric scattering model, making the restoration process more interpretable in complex real-world conditions. To bridge the synthetic-to-real gap, we design an online non-"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"PRISM achieves state-of-the-art performance on RID tasks through Proximal Scattered Atmosphere Reconstruction (PSAR), a physically structured framework that jointly reconstructs the clear scene and scattering variables under the atmospheric scattering model, combined with an online non-uniform haze synthesis pipeline and Selective Self-distillation Adaptation scheme.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the model's intrinsic scattering understanding can reliably audit residual haze and guide self-refinement in unpaired real-world scenarios without introducing new artifacts or overfitting to synthetic patterns.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"PRISM proposes a physically structured PSAR framework with non-uniform haze synthesis and selective self-distillation adaptation to achieve state-of-the-art real-world image dehazing.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"PRISM jointly reconstructs clear scenes and scattering variables to improve real-world image dehazing.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"9d6dfb958299cc02c5754d59221754e6d5a0b7c7ea259b78f87e83aba3f083f7"},"source":{"id":"2604.07048","kind":"arxiv","version":2},"verdict":{"id":"da5627d1-447d-42d8-841a-32e3924f44c9","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T18:46:11.844145Z","strongest_claim":"PRISM achieves state-of-the-art performance on RID tasks through Proximal Scattered Atmosphere Reconstruction (PSAR), a physically structured framework that jointly reconstructs the clear scene and scattering variables under the atmospheric scattering model, combined with an online non-uniform haze synthesis pipeline and Selective Self-distillation Adaptation scheme.","one_line_summary":"PRISM proposes a physically structured PSAR framework with non-uniform haze synthesis and selective self-distillation adaptation to achieve state-of-the-art real-world image dehazing.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the model's intrinsic scattering understanding can reliably audit residual haze and guide self-refinement in unpaired real-world scenarios without introducing new artifacts or overfitting to synthetic patterns.","pith_extraction_headline":"PRISM jointly reconstructs clear scenes and scattering variables to improve real-world image dehazing."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.07048/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"7125c7506fa63774af84664caded786c15e56e49a515f1c3f0d00d7d4ad87ff2"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}