logo
CanyamAI summaries for academic research
language
MobileMobile
  • Phone FeatureWeChat Mini Program
EnglishEnglish
noticeNotice
BlogBlog
delete
arrow
Return

MGCFDN: Image copy-move forgery detection method based on multi-granularity feature consistency

delete2025-11-06
delete0
PRE
AI
H
Hang Tu *
L
Liang Peng *
X
Xiaoguang Lu
赵
赵慧敏 (Huimin Zhao)
DOI：10.1016/j.neucom.2025.132029delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En
• This paper designs an end-to-end multi-granularity image copy-move forgery detection network, which performs copy-move forgery localization and source-target differentiation by extracting consistency features at various granularities. • A multi-granularity consistency module is proposed, which extracts four types of consistency features at different granularities from the image’s low-level representations, covering consistency from coarse-grained to fine-grained and from local granularity to global granularity, progressively enhancing the clarity of the copy-move forgery region’s decision boundaries. • On various structured visual backbone networks and multiple generalization datasets, our method achieves the best Precision in grid structures and the best Recall in graph structures. • By comparing with several existing advanced methods, our approach achieves higher accuracy and stronger generalization capability on five mainstream public datasets.

Journal

Neurocomputing cover
Neurocomputing
IF:
6.5
Papers:
2.5W
Citations:
6.5W

Organization

No organization information available