Lego-Edit: A General Image Editing Framework with Model-Level Bricks and MLLM Builder Article Swipe

PDF

Qiong Jia , Yu Liu , Yajie Chai , Xintong Yao , Qiming Lu , Yasen Zhang , R. S. Shi , Ying Huang , Guoquan Zhang ·

YOU? · · 2025 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2509.12883

Instruction-based image editing has garnered significant attention due to its direct interaction with users. However, real-world user instructions are immensely diverse, and existing methods often fail to generalize effectively to instructions outside their training domain, limiting their practical application. To address this, we propose Lego-Edit, which leverages the generalization capability of Multi-modal Large Language Model (MLLM) to organize a suite of model-level editing tools to tackle this challenge. Lego-Edit incorporates two key designs: (1) a model-level toolkit comprising diverse models efficiently trained on limited data and several image manipulation functions, enabling fine-grained composition of editing actions by the MLLM; and (2) a three-stage progressive reinforcement learning approach that uses feedback on unannotated, open-domain instructions to train the MLLM, equipping it with generalized reasoning capabilities for handling real-world instructions. Experiments demonstrate that Lego-Edit achieves state-of-the-art performance on GEdit-Bench and ImgBench. It exhibits robust reasoning capabilities for open-domain instructions and can utilize newly introduced editing tools without additional fine-tuning. Code is available: https://github.com/xiaomi-research/lego-edit.

Related Topics

Virginia General Assembly

The General Theory Of Employment, Interest And Money

2018 Democratic Republic Of The Congo General Election

General Dynamics F-16 Fighting Falcon Variants

2019 United Kingdom General Election

Aspect Ratio (Image)

General Lee (Car)

General Electric F414

1997 United Kingdom General Election

Public Image Of Vladimir Putin

Charles Lee (General)

General Secretary Of The Communist Party Of The Soviet Union

Concepts

No concepts available.

Metadata

Type: preprint
Language: en
Landing Page: http://arxiv.org/abs/2509.12883
PDF: https://arxiv.org/pdf/2509.12883
OA Status: green
OpenAlex ID: https://openalex.org/W4415317235

All OpenAlex metadata

Raw OpenAlex JSON

OpenAlex ID: https://openalex.org/W4415317235

Canonical identifier for this work in OpenAlex
DOI: https://doi.org/10.48550/arxiv.2509.12883

Digital Object Identifier
Title: Lego-Edit: A General Image Editing Framework with Model-Level Bricks and MLLM Builder

Work title
Type: preprint

OpenAlex work type
Language: en

Primary language
Publication year: 2025

Year of publication
Publication date: 2025-09-16

Full publication date if available
Authors: Qiong Jia, Yu Liu, Yajie Chai, Xintong Yao, Qiming Lu, Yasen Zhang, R. S. Shi, Ying Huang, Guoquan Zhang

List of authors in order
Landing page: https://arxiv.org/abs/2509.12883

Publisher landing page
PDF URL: https://arxiv.org/pdf/2509.12883

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/2509.12883

Direct OA link when available
Cited by: 0

Total citation count in OpenAlex

Full payload

id	https://openalex.org/W4415317235
doi	https://doi.org/10.48550/arxiv.2509.12883
ids.doi	https://doi.org/10.48550/arxiv.2509.12883
ids.openalex	https://openalex.org/W4415317235
fwci
type	preprint
title	Lego-Edit: A General Image Editing Framework with Model-Level Bricks and MLLM Builder
biblio.issue
biblio.volume
biblio.last_page
biblio.first_page
topics[0].id	https://openalex.org/T13067
topics[0].field.id	https://openalex.org/fields/19
topics[0].field.display_name	Earth and Planetary Sciences
topics[0].score	0.8666999936103821
topics[0].domain.id	https://openalex.org/domains/3
topics[0].domain.display_name	Physical Sciences
topics[0].subfield.id	https://openalex.org/subfields/1906
topics[0].subfield.display_name	Geochemistry and Petrology
topics[0].display_name	Geological Modeling and Analysis
topics[1].id	https://openalex.org/T10715
topics[1].field.id	https://openalex.org/fields/17
topics[1].field.display_name	Computer Science
topics[1].score	0.8568999767303467
topics[1].domain.id	https://openalex.org/domains/3
topics[1].domain.display_name	Physical Sciences
topics[1].subfield.id	https://openalex.org/subfields/1705
topics[1].subfield.display_name	Computer Networks and Communications
topics[1].display_name	Distributed and Parallel Computing Systems
topics[2].id	https://openalex.org/T12698
topics[2].field.id	https://openalex.org/fields/22
topics[2].field.display_name	Engineering
topics[2].score	0.8392000198364258
topics[2].domain.id	https://openalex.org/domains/3
topics[2].domain.display_name	Physical Sciences
topics[2].subfield.id	https://openalex.org/subfields/2215
topics[2].subfield.display_name	Building and Construction
topics[2].display_name	3D Modeling in Geospatial Applications
is_xpac	False
apc_list
apc_paid
language	en
locations[0].id	pmh:oai:arXiv.org:2509.12883
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/2509.12883
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/2509.12883
locations[1].id	doi:10.48550/arxiv.2509.12883
locations[1].is_oa	True
locations[1].source.id	https://openalex.org/S4306400194
locations[1].source.issn
locations[1].source.type	repository
locations[1].source.is_oa	True
locations[1].source.issn_l
locations[1].source.is_core	False
locations[1].source.is_in_doaj	False
locations[1].source.display_name	arXiv (Cornell University)
locations[1].source.host_organization	https://openalex.org/I205783295
locations[1].source.host_organization_name	Cornell University
locations[1].source.host_organization_lineage	https://openalex.org/I205783295
locations[1].license
locations[1].pdf_url
locations[1].version
locations[1].raw_type	article
locations[1].license_id
locations[1].is_accepted	False
locations[1].is_published
locations[1].raw_source_name
locations[1].landing_page_url	https://doi.org/10.48550/arxiv.2509.12883
indexed_in	arxiv, datacite
authorships[0].author.id	https://openalex.org/A5112913806
authorships[0].author.orcid
authorships[0].author.display_name	Qiong Jia
authorships[0].author_position	first
authorships[0].raw_author_name	Jia, Qifei
authorships[0].is_corresponding	False
authorships[1].author.id	https://openalex.org/A5004545610
authorships[1].author.orcid	https://orcid.org/0000-0002-2399-2829
authorships[1].author.display_name	Yu Liu
authorships[1].author_position	middle
authorships[1].raw_author_name	Liu, Yu
authorships[1].is_corresponding	False
authorships[2].author.id	https://openalex.org/A5032188960
authorships[2].author.orcid
authorships[2].author.display_name	Yajie Chai
authorships[2].author_position	middle
authorships[2].raw_author_name	Chai, Yajie
authorships[2].is_corresponding	False
authorships[3].author.id	https://openalex.org/A5032534745
authorships[3].author.orcid	https://orcid.org/0000-0003-2478-0278
authorships[3].author.display_name	Xintong Yao
authorships[3].author_position	middle
authorships[3].raw_author_name	Yao, Xintong
authorships[3].is_corresponding	False
authorships[4].author.id	https://openalex.org/A5101194897
authorships[4].author.orcid
authorships[4].author.display_name	Qiming Lu
authorships[4].author_position	middle
authorships[4].raw_author_name	Lu, Qiming
authorships[4].is_corresponding	False
authorships[5].author.id	https://openalex.org/A5008604152
authorships[5].author.orcid
authorships[5].author.display_name	Yasen Zhang
authorships[5].author_position	middle
authorships[5].raw_author_name	Zhang, Yasen
authorships[5].is_corresponding	False
authorships[6].author.id	https://openalex.org/A5107859614
authorships[6].author.orcid
authorships[6].author.display_name	R. S. Shi
authorships[6].author_position	middle
authorships[6].raw_author_name	Shi, Runyu
authorships[6].is_corresponding	False
authorships[7].author.id	https://openalex.org/A5019557372
authorships[7].author.orcid	https://orcid.org/0000-0002-3840-951X
authorships[7].author.display_name	Ying Huang
authorships[7].author_position	middle
authorships[7].raw_author_name	Huang, Ying
authorships[7].is_corresponding	False
authorships[8].author.id	https://openalex.org/A5100686848
authorships[8].author.orcid	https://orcid.org/0000-0002-0274-5335
authorships[8].author.display_name	Guoquan Zhang
authorships[8].author_position	last
authorships[8].raw_author_name	Zhang, Guoquan
authorships[8].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/2509.12883
open_access.oa_status	green
open_access.any_repository_has_fulltext	False
created_date	2025-10-18T00:00:00
display_name	Lego-Edit: A General Image Editing Framework with Model-Level Bricks and MLLM Builder
has_fulltext	False
is_retracted	False
updated_date	2025-11-06T06:51:31.235846
primary_topic.id	https://openalex.org/T13067
primary_topic.field.id	https://openalex.org/fields/19
primary_topic.field.display_name	Earth and Planetary Sciences
primary_topic.score	0.8666999936103821
primary_topic.domain.id	https://openalex.org/domains/3
primary_topic.domain.display_name	Physical Sciences
primary_topic.subfield.id	https://openalex.org/subfields/1906
primary_topic.subfield.display_name	Geochemistry and Petrology
primary_topic.display_name	Geological Modeling and Analysis
cited_by_count	0
locations_count	2
best_oa_location.id	pmh:oai:arXiv.org:2509.12883
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/2509.12883
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/2509.12883
primary_location.id	pmh:oai:arXiv.org:2509.12883
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/2509.12883
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/2509.12883
publication_date	2025-09-16
publication_year	2025
referenced_works_count	0
abstract_inverted_index.a	58, 74, 101
abstract_inverted_index.It	139
abstract_inverted_index.To	39
abstract_inverted_index.by	96
abstract_inverted_index.is	158
abstract_inverted_index.it	119
abstract_inverted_index.of	50, 60, 93
abstract_inverted_index.on	82, 110, 135
abstract_inverted_index.to	8, 26, 29, 56, 64, 114
abstract_inverted_index.we	42
abstract_inverted_index.(1)	73
abstract_inverted_index.(2)	100
abstract_inverted_index.and	21, 85, 99, 137, 147
abstract_inverted_index.are	18
abstract_inverted_index.can	148
abstract_inverted_index.due	7
abstract_inverted_index.for	124, 144
abstract_inverted_index.has	3
abstract_inverted_index.its	9
abstract_inverted_index.key	71
abstract_inverted_index.the	47, 97, 116
abstract_inverted_index.two	70
abstract_inverted_index.Code	157
abstract_inverted_index.data	84
abstract_inverted_index.fail	25
abstract_inverted_index.that	107, 130
abstract_inverted_index.this	66
abstract_inverted_index.user	16
abstract_inverted_index.uses	108
abstract_inverted_index.with	12, 120
abstract_inverted_index.Large	52
abstract_inverted_index.MLLM,	117
abstract_inverted_index.MLLM;	98
abstract_inverted_index.Model	54
abstract_inverted_index.image	1, 87
abstract_inverted_index.newly	150
abstract_inverted_index.often	24
abstract_inverted_index.suite	59
abstract_inverted_index.their	32, 36
abstract_inverted_index.this,	41
abstract_inverted_index.tools	63, 153
abstract_inverted_index.train	115
abstract_inverted_index.which	45
abstract_inverted_index.(MLLM)	55
abstract_inverted_index.direct	10
abstract_inverted_index.models	79
abstract_inverted_index.robust	141
abstract_inverted_index.tackle	65
abstract_inverted_index.users.	13
abstract_inverted_index.actions	95
abstract_inverted_index.address	40
abstract_inverted_index.diverse	78
abstract_inverted_index.domain,	34
abstract_inverted_index.editing	2, 62, 94, 152
abstract_inverted_index.limited	83
abstract_inverted_index.methods	23
abstract_inverted_index.outside	31
abstract_inverted_index.propose	43
abstract_inverted_index.several	86
abstract_inverted_index.toolkit	76
abstract_inverted_index.trained	81
abstract_inverted_index.utilize	149
abstract_inverted_index.without	154
abstract_inverted_index.However,	14
abstract_inverted_index.Language	53
abstract_inverted_index.achieves	132
abstract_inverted_index.approach	106
abstract_inverted_index.designs:	72
abstract_inverted_index.diverse,	20
abstract_inverted_index.enabling	90
abstract_inverted_index.exhibits	140
abstract_inverted_index.existing	22
abstract_inverted_index.feedback	109
abstract_inverted_index.garnered	4
abstract_inverted_index.handling	125
abstract_inverted_index.learning	105
abstract_inverted_index.limiting	35
abstract_inverted_index.organize	57
abstract_inverted_index.training	33
abstract_inverted_index.ImgBench.	138
abstract_inverted_index.Lego-Edit	68, 131
abstract_inverted_index.attention	6
abstract_inverted_index.equipping	118
abstract_inverted_index.immensely	19
abstract_inverted_index.leverages	46
abstract_inverted_index.practical	37
abstract_inverted_index.reasoning	122, 142
abstract_inverted_index.Lego-Edit,	44
abstract_inverted_index.additional	155
abstract_inverted_index.available:	159
abstract_inverted_index.capability	49
abstract_inverted_index.challenge.	67
abstract_inverted_index.comprising	77
abstract_inverted_index.functions,	89
abstract_inverted_index.generalize	27
abstract_inverted_index.introduced	151
abstract_inverted_index.real-world	15, 126
abstract_inverted_index.Experiments	128
abstract_inverted_index.GEdit-Bench	136
abstract_inverted_index.Multi-modal	51
abstract_inverted_index.composition	92
abstract_inverted_index.demonstrate	129
abstract_inverted_index.effectively	28
abstract_inverted_index.efficiently	80
abstract_inverted_index.generalized	121
abstract_inverted_index.interaction	11
abstract_inverted_index.model-level	61, 75
abstract_inverted_index.open-domain	112, 145
abstract_inverted_index.performance	134
abstract_inverted_index.progressive	103
abstract_inverted_index.significant	5
abstract_inverted_index.three-stage	102
abstract_inverted_index.application.	38
abstract_inverted_index.capabilities	123, 143
abstract_inverted_index.fine-grained	91
abstract_inverted_index.fine-tuning.	156
abstract_inverted_index.incorporates	69
abstract_inverted_index.instructions	17, 30, 113, 146
abstract_inverted_index.manipulation	88
abstract_inverted_index.unannotated,	111
abstract_inverted_index.instructions.	127
abstract_inverted_index.reinforcement	104
abstract_inverted_index.generalization	48
abstract_inverted_index.state-of-the-art	133
abstract_inverted_index.Instruction-based	0
abstract_inverted_index.https://github.com/xiaomi-research/lego-edit.	160
cited_by_percentile_year
countries_distinct_count	0
institutions_distinct_count	9
citation_normalized_percentile