h-Edit: Effective and Flexible Diffusion-Based Editing via Doob's h-Transform Article Swipe

PDF

Toan Nguyen , Kien Do , Duc Kieu , Thin Nguyen ·

YOU? · · 2025 · Open Access ·

We introduce a theoretical framework for diffusion-based image editing by formulating it as a reverse-time bridge modeling problem. This approach modifies the backward process of a pretrained diffusion model to construct a bridge that converges to an implicit distribution associated with the editing target at time 0. Building on this framework, we propose h-Edit, a novel editing method that utilizes Doob's h-transform and Langevin Monte Carlo to decompose the update of an intermediate edited sample into two components: a "reconstruction" term and an "editing" term. This decomposition provides flexibility, allowing the reconstruction term to be computed via existing inversion techniques and enabling the combination of multiple editing terms to handle complex editing tasks. To our knowledge, h-Edit is the first training-free method capable of performing simultaneous text-guided and reward-model-based editing. Extensive experiments, both quantitative and qualitative, show that h-Edit outperforms state-of-the-art baselines in terms of editing effectiveness and faithfulness. Our source code is available at https://github.com/nktoan/h-edit.

Related Topics

Concepts

No concepts available.

Metadata

Type: article
Language: en
Landing Page: http://arxiv.org/abs/2503.02187
PDF: https://arxiv.org/pdf/2503.02187
OA Status: green
OpenAlex ID: https://openalex.org/W4415334547

All OpenAlex metadata

Raw OpenAlex JSON

OpenAlex ID: https://openalex.org/W4415334547

Canonical identifier for this work in OpenAlex
Title: h-Edit: Effective and Flexible Diffusion-Based Editing via Doob's h-Transform

Work title
Type: article

OpenAlex work type
Language: en

Primary language
Publication year: 2025

Year of publication
Publication date: 2025-03-04

Full publication date if available
Authors: Toan Nguyen, Kien Do, Duc Kieu, Thin Nguyen

List of authors in order
Landing page: https://arxiv.org/abs/2503.02187

Publisher landing page
PDF URL: https://arxiv.org/pdf/2503.02187

Direct link to full text PDF
Open access: Yes

Whether a free full text is available
OA status: green

Open access status per OpenAlex
OA URL: https://arxiv.org/pdf/2503.02187

Direct OA link when available
Cited by: 0

Total citation count in OpenAlex

Full payload

id	https://openalex.org/W4415334547
doi
ids.openalex	https://openalex.org/W4415334547
fwci	0.0
type	article
title	h-Edit: Effective and Flexible Diffusion-Based Editing via Doob's h-Transform
biblio.issue
biblio.volume
biblio.last_page
biblio.first_page
topics[0].id	https://openalex.org/T11975
topics[0].field.id	https://openalex.org/fields/17
topics[0].field.display_name	Computer Science
topics[0].score	0.5580000281333923
topics[0].domain.id	https://openalex.org/domains/3
topics[0].domain.display_name	Physical Sciences
topics[0].subfield.id	https://openalex.org/subfields/1702
topics[0].subfield.display_name	Artificial Intelligence
topics[0].display_name	Evolutionary Algorithms and Applications
is_xpac	False
apc_list
apc_paid
language	en
locations[0].id	pmh:oai:arXiv.org:2503.02187
locations[0].is_oa	True
locations[0].source.id	https://openalex.org/S4306400194
locations[0].source.issn
locations[0].source.type	repository
locations[0].source.is_oa	True
locations[0].source.issn_l
locations[0].source.is_core	False
locations[0].source.is_in_doaj	False
locations[0].source.display_name	arXiv (Cornell University)
locations[0].source.host_organization	https://openalex.org/I205783295
locations[0].source.host_organization_name	Cornell University
locations[0].source.host_organization_lineage	https://openalex.org/I205783295
locations[0].license
locations[0].pdf_url	https://arxiv.org/pdf/2503.02187
locations[0].version	submittedVersion
locations[0].raw_type	text
locations[0].license_id
locations[0].is_accepted	False
locations[0].is_published	False
locations[0].raw_source_name
locations[0].landing_page_url	http://arxiv.org/abs/2503.02187
indexed_in	arxiv
authorships[0].author.id	https://openalex.org/A5102874067
authorships[0].author.orcid	https://orcid.org/0000-0003-2734-0622
authorships[0].author.display_name	Toan Nguyen
authorships[0].author_position	first
authorships[0].raw_author_name	Nguyen, Toan
authorships[0].is_corresponding	False
authorships[1].author.id	https://openalex.org/A5001806269
authorships[1].author.orcid	https://orcid.org/0000-0002-0119-122X
authorships[1].author.display_name	Kien Do
authorships[1].author_position	middle
authorships[1].raw_author_name	Do, Kien
authorships[1].is_corresponding	False
authorships[2].author.id	https://openalex.org/A5028613041
authorships[2].author.orcid	https://orcid.org/0009-0008-4359-3383
authorships[2].author.display_name	Duc Kieu
authorships[2].author_position	middle
authorships[2].raw_author_name	Kieu, Duc
authorships[2].is_corresponding	False
authorships[3].author.id	https://openalex.org/A5100705489
authorships[3].author.orcid	https://orcid.org/0000-0003-3467-8963
authorships[3].author.display_name	Thin Nguyen
authorships[3].author_position	last
authorships[3].raw_author_name	Nguyen, Thin
authorships[3].is_corresponding	False
has_content.pdf	False
has_content.grobid_xml	False
is_paratext	False
open_access.is_oa	True
open_access.oa_url	https://arxiv.org/pdf/2503.02187
open_access.oa_status	green
open_access.any_repository_has_fulltext	False
created_date	2025-10-19T00:00:00
display_name	h-Edit: Effective and Flexible Diffusion-Based Editing via Doob's h-Transform
has_fulltext	False
is_retracted	False
updated_date	2025-11-06T04:12:42.849631
primary_topic.id	https://openalex.org/T11975
primary_topic.field.id	https://openalex.org/fields/17
primary_topic.field.display_name	Computer Science
primary_topic.score	0.5580000281333923
primary_topic.domain.id	https://openalex.org/domains/3
primary_topic.domain.display_name	Physical Sciences
primary_topic.subfield.id	https://openalex.org/subfields/1702
primary_topic.subfield.display_name	Artificial Intelligence
primary_topic.display_name	Evolutionary Algorithms and Applications
cited_by_count	0
locations_count	1
best_oa_location.id	pmh:oai:arXiv.org:2503.02187
best_oa_location.is_oa	True
best_oa_location.source.id	https://openalex.org/S4306400194
best_oa_location.source.issn
best_oa_location.source.type	repository
best_oa_location.source.is_oa	True
best_oa_location.source.issn_l
best_oa_location.source.is_core	False
best_oa_location.source.is_in_doaj	False
best_oa_location.source.display_name	arXiv (Cornell University)
best_oa_location.source.host_organization	https://openalex.org/I205783295
best_oa_location.source.host_organization_name	Cornell University
best_oa_location.source.host_organization_lineage	https://openalex.org/I205783295
best_oa_location.license
best_oa_location.pdf_url	https://arxiv.org/pdf/2503.02187
best_oa_location.version	submittedVersion
best_oa_location.raw_type	text
best_oa_location.license_id
best_oa_location.is_accepted	False
best_oa_location.is_published	False
best_oa_location.raw_source_name
best_oa_location.landing_page_url	http://arxiv.org/abs/2503.02187
primary_location.id	pmh:oai:arXiv.org:2503.02187
primary_location.is_oa	True
primary_location.source.id	https://openalex.org/S4306400194
primary_location.source.issn
primary_location.source.type	repository
primary_location.source.is_oa	True
primary_location.source.issn_l
primary_location.source.is_core	False
primary_location.source.is_in_doaj	False
primary_location.source.display_name	arXiv (Cornell University)
primary_location.source.host_organization	https://openalex.org/I205783295
primary_location.source.host_organization_name	Cornell University
primary_location.source.host_organization_lineage	https://openalex.org/I205783295
primary_location.license
primary_location.pdf_url	https://arxiv.org/pdf/2503.02187
primary_location.version	submittedVersion
primary_location.raw_type	text
primary_location.license_id
primary_location.is_accepted	False
primary_location.is_published	False
primary_location.raw_source_name
primary_location.landing_page_url	http://arxiv.org/abs/2503.02187
publication_date	2025-03-04
publication_year	2025
referenced_works_count	0
abstract_inverted_index.a	2, 13, 25, 31, 54, 78
abstract_inverted_index.0.	46
abstract_inverted_index.To	113
abstract_inverted_index.We	0
abstract_inverted_index.an	36, 71, 82
abstract_inverted_index.as	12
abstract_inverted_index.at	44, 154
abstract_inverted_index.be	94
abstract_inverted_index.by	9
abstract_inverted_index.in	142
abstract_inverted_index.is	117, 152
abstract_inverted_index.it	11
abstract_inverted_index.of	24, 70, 104, 123, 144
abstract_inverted_index.on	48
abstract_inverted_index.to	29, 35, 66, 93, 108
abstract_inverted_index.we	51
abstract_inverted_index.Our	149
abstract_inverted_index.and	62, 81, 100, 127, 134, 147
abstract_inverted_index.for	5
abstract_inverted_index.our	114
abstract_inverted_index.the	21, 41, 68, 90, 102, 118
abstract_inverted_index.two	76
abstract_inverted_index.via	96
abstract_inverted_index.This	18, 85
abstract_inverted_index.both	132
abstract_inverted_index.code	151
abstract_inverted_index.into	75
abstract_inverted_index.show	136
abstract_inverted_index.term	80, 92
abstract_inverted_index.that	33, 58, 137
abstract_inverted_index.this	49
abstract_inverted_index.time	45
abstract_inverted_index.with	40
abstract_inverted_index.Carlo	65
abstract_inverted_index.Monte	64
abstract_inverted_index.first	119
abstract_inverted_index.image	7
abstract_inverted_index.model	28
abstract_inverted_index.novel	55
abstract_inverted_index.term.	84
abstract_inverted_index.terms	107, 143
abstract_inverted_index.Doob's	60
abstract_inverted_index.bridge	15, 32
abstract_inverted_index.edited	73
abstract_inverted_index.h-Edit	116, 138
abstract_inverted_index.handle	109
abstract_inverted_index.method	57, 121
abstract_inverted_index.sample	74
abstract_inverted_index.source	150
abstract_inverted_index.target	43
abstract_inverted_index.tasks.	112
abstract_inverted_index.update	69
abstract_inverted_index.capable	122
abstract_inverted_index.complex	110
abstract_inverted_index.editing	8, 42, 56, 106, 111, 145
abstract_inverted_index.h-Edit,	53
abstract_inverted_index.process	23
abstract_inverted_index.propose	52
abstract_inverted_index.Building	47
abstract_inverted_index.Langevin	63
abstract_inverted_index.allowing	89
abstract_inverted_index.approach	19
abstract_inverted_index.backward	22
abstract_inverted_index.computed	95
abstract_inverted_index.editing.	129
abstract_inverted_index.enabling	101
abstract_inverted_index.existing	97
abstract_inverted_index.implicit	37
abstract_inverted_index.modeling	16
abstract_inverted_index.modifies	20
abstract_inverted_index.multiple	105
abstract_inverted_index.problem.	17
abstract_inverted_index.provides	87
abstract_inverted_index.utilizes	59
abstract_inverted_index."editing"	83
abstract_inverted_index.Extensive	130
abstract_inverted_index.available	153
abstract_inverted_index.baselines	141
abstract_inverted_index.construct	30
abstract_inverted_index.converges	34
abstract_inverted_index.decompose	67
abstract_inverted_index.diffusion	27
abstract_inverted_index.framework	4
abstract_inverted_index.introduce	1
abstract_inverted_index.inversion	98
abstract_inverted_index.associated	39
abstract_inverted_index.framework,	50
abstract_inverted_index.knowledge,	115
abstract_inverted_index.performing	124
abstract_inverted_index.pretrained	26
abstract_inverted_index.techniques	99
abstract_inverted_index.combination	103
abstract_inverted_index.components:	77
abstract_inverted_index.formulating	10
abstract_inverted_index.h-transform	61
abstract_inverted_index.outperforms	139
abstract_inverted_index.text-guided	126
abstract_inverted_index.theoretical	3
abstract_inverted_index.distribution	38
abstract_inverted_index.experiments,	131
abstract_inverted_index.flexibility,	88
abstract_inverted_index.intermediate	72
abstract_inverted_index.qualitative,	135
abstract_inverted_index.quantitative	133
abstract_inverted_index.reverse-time	14
abstract_inverted_index.simultaneous	125
abstract_inverted_index.decomposition	86
abstract_inverted_index.effectiveness	146
abstract_inverted_index.faithfulness.	148
abstract_inverted_index.training-free	120
abstract_inverted_index.reconstruction	91
abstract_inverted_index.diffusion-based	6
abstract_inverted_index."reconstruction"	79
abstract_inverted_index.state-of-the-art	140
abstract_inverted_index.reward-model-based	128
abstract_inverted_index.https://github.com/nktoan/h-edit.	155
cited_by_percentile_year
countries_distinct_count	0
institutions_distinct_count	4
citation_normalized_percentile.value	0.22638681
citation_normalized_percentile.is_in_top_1_percent	False
citation_normalized_percentile.is_in_top_10_percent	True