{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2606.26602", "kind": "paper", "url": "https://arxiv.org/abs/2606.26602"}, {"id": "gh:pku-icst-mipl/dico_bench_eccv2026", "kind": "repo", "url": "https://github.com/PKU-ICST-MIPL/DICO_Bench_ECCV2026"}], "authors": ["Geng Li", "Yuxin Peng"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.26602", "source_url": "https://arxiv.org/abs/2606.26602"}, "categories": ["General AI", "Vision & 3D", "cs.CV"], "description": {"en": "DiCoBench evaluates multimodal LLMs on multi-image fine-grained perception using high-resolution (up to 2K) image pairs. It includes 765 multiple-choice questions across two tracks: differential and commonality visual cues, covering 8 perception tasks, with exact-match scoring."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.26602", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.26602", "title": "DiCoBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2606.26602", "modality": null, "name": "DiCoBench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "High", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "unknown", "origin_record_id": "bm_dicobench_558e401f", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "95a63830851e861737506594fa8cdc5be3fd2cc41cf5122cf32a76b0f24d7ba3", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2606.26602"}, "publisher": null, "released": "2026-06-25", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.26602", "source_url": "https://arxiv.org/abs/2606.26602"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-26602", "source": "claire_radar", "source_benchmark_id": "2606.26602", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Vision & 3D", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/PKU-ICST-MIPL/DICO_Bench_ECCV2026", "githubScope": "benchmark_repo", "githubStars": 0, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 1, "hfPaperUrl": "https://huggingface.co/papers/2606.26602", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-10-04", "state": "stale"}}, "status": "fresh"}, "capabilities": [], "confidence": "High", "construction": "Transform Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "057c225e8c3eaf519674e7e2633fffa5561c01c21c064990aba6ff97ba03bc8e", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "DiCoBench evaluates multimodal LLMs on multi-image fine-grained perception using high-resolution (up to 2K) image pairs. It includes 765 multiple-choice questions across two tracks: differential and commonality visual cues, covering 8 perception tasks, with exact-match scoring.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "To bridge this gap, we introduce DiCoBench, a comprehensive, multi-image high-resolution benchmark designed for cross-image fine-grained perception."}, "familyId": "bmf_1608a2621237", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [{"evidence": [{"basis": "existing-code-link", "linkedUrl": "https://github.com/PKU-ICST-MIPL/DICO_Bench_ECCV2026", "url": "https://github.com/PKU-ICST-MIPL/DICO_Bench_ECCV2026"}, {"basis": "existing-metric-association", "linkedUrl": "https://github.com/PKU-ICST-MIPL/DICO_Bench_ECCV2026", "url": "https://huggingface.co/papers/2606.26602"}, {"basis": "linked-from-report", "linkedUrl": "https://github.com/PKU-ICST-MIPL/DICO_Bench_ECCV2026", "url": "https://arxiv.org/abs/2606.26602"}], "scope": "benchmark_repo", "url": "https://github.com/PKU-ICST-MIPL/DICO_Bench_ECCV2026", "verification": "source-linked"}], "status": "linked"}, "heat": null, "id": "bm_dicobench_558e401f", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": "https://github.com/PKU-ICST-MIPL/DICO_Bench_ECCV2026", "data": null, "hfPaper": "https://huggingface.co/papers/2606.26602", "paper": "https://arxiv.org/abs/2606.26602", "pdf": "https://arxiv.org/pdf/2606.26602", "project": null, "report": "https://arxiv.org/abs/2606.26602"}, "metrics": [], "motivation": "Recent advancements in Multimodal Large Language Models (MLLMs) have demonstrated impressive fine-grained perception capabilities.", "name": "DiCoBench", "oneLine": "DiCoBench evaluates multimodal LLMs on multi-image fine-grained perception using high-resolution (up to 2K) image pairs. It includes 765 multiple-choice questions across two tracks: differential and commonality visual cues, covering 8 perception tasks, with exact-match scoring.", "paperTitle": "DiCoBench: Benchmarking Multi-Image Fine-Grained Perception via Differential and Commonality Visual Cues", "primaryDomain": "General AI", "publication": {"evidence": "Accepted by ECCV 2026. Project page with code: https://github.com/PKU-ICST-MIPL/DICO_Bench_ECCV2026", "evidenceLevel": "author-claim", "evidenceUrl": "https://arxiv.org/abs/2606.26602", "source": "arxiv-comments", "status": "acceptance_claimed", "venue": "ECCV 2026", "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "PKU-ICST-MIPL", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://github.com/PKU-ICST-MIPL/DICO_Bench_ECCV2026"}], "ranking": {}, "readiness": "Runnable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-06-25", "source": {"authors": ["Geng Li", "Yuxin Peng"], "categories": ["cs.CV"], "id": "2606.26602", "title": "DiCoBench: Benchmarking Multi-Image Fine-Grained Perception via Differential and Commonality Visual Cues", "type": "arxiv", "url": "https://arxiv.org/abs/2606.26602"}, "sourceUpdatedAt": "2026-06-25T05:02:38Z", "topics": ["cs.CV"], "venueAttempts": [{"decisionRaw": "Accepted by ECCV 2026. Project page with code: https://github.com/PKU-ICST-MIPL/DICO_Bench_ECCV2026", "evidence": [{"level": "author-claim", "observedAt": "2026-10-14T01:46:58.058199Z", "rawValue": "Accepted by ECCV 2026. Project page with code: https://github.com/PKU-ICST-MIPL/DICO_Bench_ECCV2026", "sourceType": "arxiv-comments", "sourceUrl": "https://arxiv.org/abs/2606.26602"}], "reviewStatus": "accepted", "venueName": "ECCV 2026"}], "whyItMatters": "Existing benchmarks rely on explicit textual cues or low resolution, whereas DiCoBench targets autonomous discovery of subtle visual cues in high-resolution pairs. It provides a challenging testbed with a large human-model performance gap, aiding progress in complex multi-image understanding."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
