{"paper":{"title":"Venus-DeFakerOne: Unified Fake Image Detection & Localization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"DeFakerOne integrates InternVL2 and SAM2 into one model that detects and localizes image forgeries across many generation types at once.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"GuangJian Team","submitted_at":"2026-05-13T20:20:33Z","abstract_excerpt":"In recent years, the rapid evolution of generative AI has fundamentally reshaped the paradigm of image forgery, breaking the traditional boundaries between document editing, natural image manipulation, DeepFake generation, and full-image AIGC synthesis. Despite this shift toward unified forgery generation, existing research in Fake Image Detection and Localization (FIDL) remains fragmented. This creates a mismatch between increasingly unified forgery generation mechanisms and the domain-specific detection paradigm. Bridging this mismatch poses two key challenges for FIDL: understanding cross-d"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"DeFakerOne achieves state-of-the-art performance, outperforming baselines on 39 forgery detection benchmarks and 9 localization benchmarks.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That integrating InternVL2 and SAM2 with fine-grained supervision sufficiently captures cross-domain artifact transfer and interference without domain-specific adaptations.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"DeFakerOne integrates InternVL2 and SAM2 into a single model that achieves state-of-the-art results on 39 detection and 9 localization benchmarks for unified fake image detection and localization.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"DeFakerOne integrates InternVL2 and SAM2 into one model that detects and localizes image forgeries across many generation types at once.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"ed973a2587b9722d376254be1c3e5915b5bd0b7e736be819e5f6222315ae2f62"},"source":{"id":"2605.14091","kind":"arxiv","version":1},"verdict":{"id":"8e319d0d-7138-4122-9a47-557cd0be4388","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T05:25:41.568710Z","strongest_claim":"DeFakerOne achieves state-of-the-art performance, outperforming baselines on 39 forgery detection benchmarks and 9 localization benchmarks.","one_line_summary":"DeFakerOne integrates InternVL2 and SAM2 into a single model that achieves state-of-the-art results on 39 detection and 9 localization benchmarks for unified fake image detection and localization.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That integrating InternVL2 and SAM2 with fine-grained supervision sufficiently captures cross-domain artifact transfer and interference without domain-specific adaptations.","pith_extraction_headline":"DeFakerOne integrates InternVL2 and SAM2 into one model that detects and localizes image forgeries across many generation types at once."},"references":{"count":240,"sample":[{"doi":"","year":2025,"title":"Findings of the Association for Computational Linguistics: EMNLP 2025 , pages=","work_id":"6bb81f2e-0894-4d04-ac96-0b406a06140a","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Ivy-Fake: A Unified Explainable Framework and Benchmark for Image and Video AIGC Detection","work_id":"f28a8b39-972c-46e1-85e2-94c11302493e","ref_index":2,"cited_arxiv_id":"2506.00979","is_internal_anchor":true},{"doi":"","year":null,"title":"Fake-HR1: Rethinking Reasoning of Vision Language Model for Synthetic Image Detection , year=","work_id":"b3c836c1-b202-4f85-ad2b-0fe08c4b4978","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"The Fourteenth International Conference on Learning Representations , year=","work_id":"e8971a5c-107c-48b8-908f-1ad1f40e577d","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Effort: Efficient Orthogonal Modeling for Generalizable AI-Generated Image Detection , author=. ICML , year=","work_id":"7f1d763b-c6a7-4280-b3ac-0c5964af5e86","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":240,"snapshot_sha256":"3c6447ba8f4a3dec70ab7a9b6df50a21d826b3fd666c20097b76af70c58198ba","internal_anchors":13},"formal_canon":{"evidence_count":2,"snapshot_sha256":"36ca793253c2e3d07f53b691468c3fb3817fd8be953a450785426c16a88d27b7"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}