{"paper":{"title":"DSH-Bench: A Difficulty- and Scenario-Aware Benchmark with Hierarchical Subject Taxonomy for Subject-Driven Text-to-Image Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"DSH-Bench supplies a hierarchical taxonomy and difficulty-scenario labels to expose where subject-driven text-to-image models lose identity.","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Chao Deng, Hang Chen, Huan Yu, Jie Jiang, Liqun Liu, Longfei Lu, Luo Liao, Mengge Xue, Peng Shu, Qing Wang, Shuang Li, Te Cao, Yuan Chen, Zhenyu Hu","submitted_at":"2026-03-09T08:30:28Z","abstract_excerpt":"Significant progress has been achieved in subject-driven text-to-image (T2I) generation, which aims to synthesize new images depicting target subjects according to user instructions. However, evaluating these models remains a significant challenge. Existing benchmarks exhibit critical limitations: 1) insufficient diversity and comprehensiveness in subject images, 2) inadequate granularity in assessing model performance across different subject difficulty levels and prompt scenarios, and 3) a profound lack of actionable insights and diagnostic guidance for subsequent model refinement. To addres"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"DSH-Bench enables systematic multi-perspective analysis of subject-driven T2I models through four principal innovations: hierarchical taxonomy sampling, difficulty and scenario classification, the SICS metric with 9.4% higher human correlation, and diagnostic insights for model optimization.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that the chosen hierarchical taxonomy and difficulty/scenario labels are sufficiently comprehensive and unbiased, and that the reported 9.4% correlation improvement for SICS generalizes beyond the specific human evaluators and model set used in the study.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"DSH-Bench is a benchmark for subject-driven T2I generation that uses hierarchical taxonomy sampling, difficulty/scenario classification, and a new SICS metric showing 9.4% higher human correlation than prior measures.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"DSH-Bench supplies a hierarchical taxonomy and difficulty-scenario labels to expose where subject-driven text-to-image models lose identity.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"2856dc4eb7249d7c68e586f81f744d7bcf9d09dc39de0ecdac832c929c77bf76"},"source":{"id":"2603.08090","kind":"arxiv","version":3},"verdict":{"id":"ce162a79-d45e-42c0-8f1d-cc6f950a0c40","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T15:13:21.785111Z","strongest_claim":"DSH-Bench enables systematic multi-perspective analysis of subject-driven T2I models through four principal innovations: hierarchical taxonomy sampling, difficulty and scenario classification, the SICS metric with 9.4% higher human correlation, and diagnostic insights for model optimization.","one_line_summary":"DSH-Bench is a benchmark for subject-driven T2I generation that uses hierarchical taxonomy sampling, difficulty/scenario classification, and a new SICS metric showing 9.4% higher human correlation than prior measures.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that the chosen hierarchical taxonomy and difficulty/scenario labels are sufficiently comprehensive and unbiased, and that the reported 9.4% correlation improvement for SICS generalizes beyond the specific human evaluators and model set used in the study.","pith_extraction_headline":"DSH-Bench supplies a hierarchical taxonomy and difficulty-scenario labels to expose where subject-driven text-to-image models lose identity."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2603.08090/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":"866d851962e16abf12a6f4c6d82bc8bb7da7fed262412480a65abef9e8bef40b"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}