Hybrid Domain Perception Combined With Multi-expert Decoding to Improve Image Forgery Localization
Abstract With the advancements in multimedia software and hardware technology, image forgery localization has become an important challenge in digital forensics. To improve the efficiency and stability of image forgery detection, we propose a mixed-domain perception and multi-expert decoding recogni...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2025-07-01
|
| Series: | International Journal of Computational Intelligence Systems |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s44196-025-00892-7 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849331825988599808 |
|---|---|
| author | Xuchao Gong Hongjie Duan Peiying Zhang Jian Wang Kai Liu Zhaohui Li |
| author_facet | Xuchao Gong Hongjie Duan Peiying Zhang Jian Wang Kai Liu Zhaohui Li |
| author_sort | Xuchao Gong |
| collection | DOAJ |
| description | Abstract With the advancements in multimedia software and hardware technology, image forgery localization has become an important challenge in digital forensics. To improve the efficiency and stability of image forgery detection, we propose a mixed-domain perception and multi-expert decoding recognition model. First, we design an alignment strategy that utilizes both RGB and frequency domain information of images. This strategy adapts to the multi-dimensional distribution characteristics of the original data, enhancing the discrimination of tampered regions. Next, we employ a hybrid expert modeling approach to improve the model’s robustness in the representation space through feature selection and recombination. Additionally, we introduce a region-weighted contrastive learning method to better localize and focus on tampered regions. Experiments on four datasets (CASIA, NIST, COVERAGE, and IMD) show that our proposed model achieves an improvement in AUC ranging from 0.15 to 1.9% compared to the existing advanced methods. These results indicate that our approach contributes to more accurate image forgery localization, offering potential benefits for digital forensics and multimedia security applications. |
| format | Article |
| id | doaj-art-65ec409dc58e4ef0839f61ea316c6109 |
| institution | Kabale University |
| issn | 1875-6883 |
| language | English |
| publishDate | 2025-07-01 |
| publisher | Springer |
| record_format | Article |
| series | International Journal of Computational Intelligence Systems |
| spelling | doaj-art-65ec409dc58e4ef0839f61ea316c61092025-08-20T03:46:24ZengSpringerInternational Journal of Computational Intelligence Systems1875-68832025-07-0118111510.1007/s44196-025-00892-7Hybrid Domain Perception Combined With Multi-expert Decoding to Improve Image Forgery LocalizationXuchao Gong0Hongjie Duan1Peiying Zhang2Jian Wang3Kai Liu4Zhaohui Li5Sinopec Shengli Oilfield BranchSinopec Shengli Oilfield BranchQingdao Institute of Software, College of Computer Science and Technology, China University of Petroleum (East China)College of Science, China University of Petroleum (East China)Beijing National Research Center for Information Science and Technology, Tsinghua UniversityCollege of Electronics and Information Engineering, South-Central Minzu UniversityAbstract With the advancements in multimedia software and hardware technology, image forgery localization has become an important challenge in digital forensics. To improve the efficiency and stability of image forgery detection, we propose a mixed-domain perception and multi-expert decoding recognition model. First, we design an alignment strategy that utilizes both RGB and frequency domain information of images. This strategy adapts to the multi-dimensional distribution characteristics of the original data, enhancing the discrimination of tampered regions. Next, we employ a hybrid expert modeling approach to improve the model’s robustness in the representation space through feature selection and recombination. Additionally, we introduce a region-weighted contrastive learning method to better localize and focus on tampered regions. Experiments on four datasets (CASIA, NIST, COVERAGE, and IMD) show that our proposed model achieves an improvement in AUC ranging from 0.15 to 1.9% compared to the existing advanced methods. These results indicate that our approach contributes to more accurate image forgery localization, offering potential benefits for digital forensics and multimedia security applications.https://doi.org/10.1007/s44196-025-00892-7Image forgery localizationInformation alignmentMixed expertsRegional weightingComparative learning |
| spellingShingle | Xuchao Gong Hongjie Duan Peiying Zhang Jian Wang Kai Liu Zhaohui Li Hybrid Domain Perception Combined With Multi-expert Decoding to Improve Image Forgery Localization International Journal of Computational Intelligence Systems Image forgery localization Information alignment Mixed experts Regional weighting Comparative learning |
| title | Hybrid Domain Perception Combined With Multi-expert Decoding to Improve Image Forgery Localization |
| title_full | Hybrid Domain Perception Combined With Multi-expert Decoding to Improve Image Forgery Localization |
| title_fullStr | Hybrid Domain Perception Combined With Multi-expert Decoding to Improve Image Forgery Localization |
| title_full_unstemmed | Hybrid Domain Perception Combined With Multi-expert Decoding to Improve Image Forgery Localization |
| title_short | Hybrid Domain Perception Combined With Multi-expert Decoding to Improve Image Forgery Localization |
| title_sort | hybrid domain perception combined with multi expert decoding to improve image forgery localization |
| topic | Image forgery localization Information alignment Mixed experts Regional weighting Comparative learning |
| url | https://doi.org/10.1007/s44196-025-00892-7 |
| work_keys_str_mv | AT xuchaogong hybriddomainperceptioncombinedwithmultiexpertdecodingtoimproveimageforgerylocalization AT hongjieduan hybriddomainperceptioncombinedwithmultiexpertdecodingtoimproveimageforgerylocalization AT peiyingzhang hybriddomainperceptioncombinedwithmultiexpertdecodingtoimproveimageforgerylocalization AT jianwang hybriddomainperceptioncombinedwithmultiexpertdecodingtoimproveimageforgerylocalization AT kailiu hybriddomainperceptioncombinedwithmultiexpertdecodingtoimproveimageforgerylocalization AT zhaohuili hybriddomainperceptioncombinedwithmultiexpertdecodingtoimproveimageforgerylocalization |