{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:53DW6HKKV5H72KZX5FGLXDBWZT","short_pith_number":"pith:53DW6HKK","schema_version":"1.0","canonical_sha256":"eec76f1d4aaf4ffd2b37e94cbb8c36cce6888f8cbcefe546cd266c26c4c9908f","source":{"kind":"arxiv","id":"2405.08246","version":1},"attestation_state":"computed","paper":{"title":"Compositional Text-to-Image Generation with Dense Blob Representations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Arash Vahdat, Benjamin Eckart, Chao Liu, Morteza Mardani, Sifei Liu, Weili Nie","submitted_at":"2024-05-14T00:22:06Z","abstract_excerpt":"Existing text-to-image models struggle to follow complex text prompts, raising the need for extra grounding inputs for better controllability. In this work, we propose to decompose a scene into visual primitives - denoted as dense blob representations - that contain fine-grained details of the scene while being modular, human-interpretable, and easy-to-construct. Based on blob representations, we develop a blob-grounded text-to-image diffusion model, termed BlobGEN, for compositional generation. Particularly, we introduce a new masked cross-attention module to disentangle the fusion between bl"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2405.08246","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-05-14T00:22:06Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"a1d74a944199bb4ac93037a3abc1b0f1a8996b82105858a22103d34502179dd0","abstract_canon_sha256":"ab0a5aab82983aee8321f3ada2e93436708f7b62df54d64d3f5f1d5a1db8e275"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:18:57.555125Z","signature_b64":"e023FHcUWPY40AVMTlLoHWsZV/Dl7v0+s+t+ytyKOfIXCOCwUNkVnk/h2fLKtmPYGnn1OUlAZxDJvMy8hsa+Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eec76f1d4aaf4ffd2b37e94cbb8c36cce6888f8cbcefe546cd266c26c4c9908f","last_reissued_at":"2026-07-05T08:18:57.554634Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:18:57.554634Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Compositional Text-to-Image Generation with Dense Blob Representations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Arash Vahdat, Benjamin Eckart, Chao Liu, Morteza Mardani, Sifei Liu, Weili Nie","submitted_at":"2024-05-14T00:22:06Z","abstract_excerpt":"Existing text-to-image models struggle to follow complex text prompts, raising the need for extra grounding inputs for better controllability. In this work, we propose to decompose a scene into visual primitives - denoted as dense blob representations - that contain fine-grained details of the scene while being modular, human-interpretable, and easy-to-construct. Based on blob representations, we develop a blob-grounded text-to-image diffusion model, termed BlobGEN, for compositional generation. Particularly, we introduce a new masked cross-attention module to disentangle the fusion between bl"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.08246","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2405.08246/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":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2405.08246","created_at":"2026-07-05T08:18:57.554694+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.08246v1","created_at":"2026-07-05T08:18:57.554694+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.08246","created_at":"2026-07-05T08:18:57.554694+00:00"},{"alias_kind":"pith_short_12","alias_value":"53DW6HKKV5H7","created_at":"2026-07-05T08:18:57.554694+00:00"},{"alias_kind":"pith_short_16","alias_value":"53DW6HKKV5H72KZX","created_at":"2026-07-05T08:18:57.554694+00:00"},{"alias_kind":"pith_short_8","alias_value":"53DW6HKK","created_at":"2026-07-05T08:18:57.554694+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.15181","citing_title":"From Plans to Pixels: Learning to Plan and Orchestrate for Open-Ended Image Editing","ref_index":30,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/53DW6HKKV5H72KZX5FGLXDBWZT","json":"https://pith.science/pith/53DW6HKKV5H72KZX5FGLXDBWZT.json","graph_json":"https://pith.science/api/pith-number/53DW6HKKV5H72KZX5FGLXDBWZT/graph.json","events_json":"https://pith.science/api/pith-number/53DW6HKKV5H72KZX5FGLXDBWZT/events.json","paper":"https://pith.science/paper/53DW6HKK"},"agent_actions":{"view_html":"https://pith.science/pith/53DW6HKKV5H72KZX5FGLXDBWZT","download_json":"https://pith.science/pith/53DW6HKKV5H72KZX5FGLXDBWZT.json","view_paper":"https://pith.science/paper/53DW6HKK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.08246&json=true","fetch_graph":"https://pith.science/api/pith-number/53DW6HKKV5H72KZX5FGLXDBWZT/graph.json","fetch_events":"https://pith.science/api/pith-number/53DW6HKKV5H72KZX5FGLXDBWZT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/53DW6HKKV5H72KZX5FGLXDBWZT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/53DW6HKKV5H72KZX5FGLXDBWZT/action/storage_attestation","attest_author":"https://pith.science/pith/53DW6HKKV5H72KZX5FGLXDBWZT/action/author_attestation","sign_citation":"https://pith.science/pith/53DW6HKKV5H72KZX5FGLXDBWZT/action/citation_signature","submit_replication":"https://pith.science/pith/53DW6HKKV5H72KZX5FGLXDBWZT/action/replication_record"}},"created_at":"2026-07-05T08:18:57.554694+00:00","updated_at":"2026-07-05T08:18:57.554694+00:00"}