{"paper":{"title":"OpenVTON-Bench: A Large-Scale High-Resolution Benchmark for Controllable Virtual Try-On Evaluation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"OpenVTON-Bench introduces a 100K-pair dataset and five-dimension protocol that correlates 0.833 with human VTON judgments.","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Chenhui Wu, Jingwen Luo, Jin Li, Kai Wen, Shuai Jiang, Siqi Yin, Tao Chen, Weijie Wang","submitted_at":"2026-01-30T08:58:00Z","abstract_excerpt":"Recent advances in diffusion models have significantly elevated the visual fidelity of Virtual Try-On (VTON) systems, yet reliable evaluation remains a persistent bottleneck. Traditional metrics struggle to quantify fine-grained texture details and semantic consistency, while existing datasets fail to meet commercial standards in scale and diversity. We present OpenVTON-Bench, a large-scale benchmark comprising approximately 100K high-resolution image pairs (up to $1536 \\times 1536$). The dataset is constructed using DINOv3-based hierarchical clustering for semantically balanced sampling and G"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experimental results show strong agreement with human judgments (Kendall's τ of 0.833 vs. 0.611 for SSIM), establishing a robust benchmark for VTON evaluation.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the VLM-based semantic reasoning combined with the Multi-Scale Representation Metric based on SAM3 segmentation and morphological erosion accurately and without bias measures the five dimensions of VTON quality.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"OpenVTON-Bench supplies a large-scale high-resolution dataset and multi-modal protocol for virtual try-on evaluation that correlates more strongly with human judgments than standard metrics like SSIM.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"OpenVTON-Bench introduces a 100K-pair dataset and five-dimension protocol that correlates 0.833 with human VTON judgments.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"ad857c3c766956369858fe77661d09771cc11962c3b8dcb68f7c53171b6d9c8f"},"source":{"id":"2601.22725","kind":"arxiv","version":4},"verdict":{"id":"6c9ffc3d-7792-4c2d-801b-16d49faeccb0","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T09:44:52.030002Z","strongest_claim":"Experimental results show strong agreement with human judgments (Kendall's τ of 0.833 vs. 0.611 for SSIM), establishing a robust benchmark for VTON evaluation.","one_line_summary":"OpenVTON-Bench supplies a large-scale high-resolution dataset and multi-modal protocol for virtual try-on evaluation that correlates more strongly with human judgments than standard metrics like SSIM.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the VLM-based semantic reasoning combined with the Multi-Scale Representation Metric based on SAM3 segmentation and morphological erosion accurately and without bias measures the five dimensions of VTON quality.","pith_extraction_headline":"OpenVTON-Bench introduces a 100K-pair dataset and five-dimension protocol that correlates 0.833 with human VTON judgments."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2601.22725/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":"7226196ab3f7a2ddf4d1d3c5e8e92af1d311906687f81573cf8ca4727afa6af2"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}