{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:3R5K6GCCEZTFQI4Z3YKTLWILQD","short_pith_number":"pith:3R5K6GCC","schema_version":"1.0","canonical_sha256":"dc7aaf18422666582399de1535d90b80cf5d497aba38912d9bdf523e2243e2a4","source":{"kind":"arxiv","id":"2303.15403","version":2},"attestation_state":"computed","paper":{"title":"Training-free Content Injection using h-space in Diffusion Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jaeseok Jeong, Mingi Kwon, Youngjung Uh","submitted_at":"2023-03-27T17:19:50Z","abstract_excerpt":"Diffusion models (DMs) synthesize high-quality images in various domains. However, controlling their generative process is still hazy because the intermediate variables in the process are not rigorously studied. Recently, the bottleneck feature of the U-Net, namely $h$-space, is found to convey the semantics of the resulting image. It enables StyleCLIP-like latent editing within DMs. In this paper, we explore further usage of $h$-space beyond attribute editing, and introduce a method to inject the content of one image into another image by combining their features in the generative processes. "},"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":"2303.15403","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2023-03-27T17:19:50Z","cross_cats_sorted":[],"title_canon_sha256":"841b94e820c38aaaada858243b1aa13632c3dabc88086d6b3732cc8bbddece60","abstract_canon_sha256":"bd7bb446dfac2c7517a6de6182057f9080ecfa381e5b78f618915591e90fce48"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:30:02.229250Z","signature_b64":"8rLxbqE3j6gHVT9d46T3GX8pU/qV+NlJ10L5KIQOsr/Ql0D2787bawvZEh7zSg9ixzePCq4jPsSP6UF3khSQAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dc7aaf18422666582399de1535d90b80cf5d497aba38912d9bdf523e2243e2a4","last_reissued_at":"2026-07-05T07:30:02.228913Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:30:02.228913Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Training-free Content Injection using h-space in Diffusion Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jaeseok Jeong, Mingi Kwon, Youngjung Uh","submitted_at":"2023-03-27T17:19:50Z","abstract_excerpt":"Diffusion models (DMs) synthesize high-quality images in various domains. However, controlling their generative process is still hazy because the intermediate variables in the process are not rigorously studied. Recently, the bottleneck feature of the U-Net, namely $h$-space, is found to convey the semantics of the resulting image. It enables StyleCLIP-like latent editing within DMs. In this paper, we explore further usage of $h$-space beyond attribute editing, and introduce a method to inject the content of one image into another image by combining their features in the generative processes. "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.15403","kind":"arxiv","version":2},"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/2303.15403/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":"2303.15403","created_at":"2026-07-05T07:30:02.228966+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.15403v2","created_at":"2026-07-05T07:30:02.228966+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.15403","created_at":"2026-07-05T07:30:02.228966+00:00"},{"alias_kind":"pith_short_12","alias_value":"3R5K6GCCEZTF","created_at":"2026-07-05T07:30:02.228966+00:00"},{"alias_kind":"pith_short_16","alias_value":"3R5K6GCCEZTFQI4Z","created_at":"2026-07-05T07:30:02.228966+00:00"},{"alias_kind":"pith_short_8","alias_value":"3R5K6GCC","created_at":"2026-07-05T07:30:02.228966+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.18251","citing_title":"Style-Based Neural Architectures for Real-Time Weather Classification","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12281","citing_title":"MAST: Mask-Guided Attention Mass Allocation for Training-Free Multi-Style Transfer","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08364","citing_title":"MegaStyle: Constructing Diverse and Scalable Style Dataset via Consistent Text-to-Image Style Mapping","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3R5K6GCCEZTFQI4Z3YKTLWILQD","json":"https://pith.science/pith/3R5K6GCCEZTFQI4Z3YKTLWILQD.json","graph_json":"https://pith.science/api/pith-number/3R5K6GCCEZTFQI4Z3YKTLWILQD/graph.json","events_json":"https://pith.science/api/pith-number/3R5K6GCCEZTFQI4Z3YKTLWILQD/events.json","paper":"https://pith.science/paper/3R5K6GCC"},"agent_actions":{"view_html":"https://pith.science/pith/3R5K6GCCEZTFQI4Z3YKTLWILQD","download_json":"https://pith.science/pith/3R5K6GCCEZTFQI4Z3YKTLWILQD.json","view_paper":"https://pith.science/paper/3R5K6GCC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.15403&json=true","fetch_graph":"https://pith.science/api/pith-number/3R5K6GCCEZTFQI4Z3YKTLWILQD/graph.json","fetch_events":"https://pith.science/api/pith-number/3R5K6GCCEZTFQI4Z3YKTLWILQD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3R5K6GCCEZTFQI4Z3YKTLWILQD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3R5K6GCCEZTFQI4Z3YKTLWILQD/action/storage_attestation","attest_author":"https://pith.science/pith/3R5K6GCCEZTFQI4Z3YKTLWILQD/action/author_attestation","sign_citation":"https://pith.science/pith/3R5K6GCCEZTFQI4Z3YKTLWILQD/action/citation_signature","submit_replication":"https://pith.science/pith/3R5K6GCCEZTFQI4Z3YKTLWILQD/action/replication_record"}},"created_at":"2026-07-05T07:30:02.228966+00:00","updated_at":"2026-07-05T07:30:02.228966+00:00"}