Cross-stream attention enhanced central difference convolutional network for CG image detection
With the maturation of computer graphics (CG) technology in the field of image generation, the realism of created images has been improved significantly. Although these technologies are widely used in daily life and bring many conveniences, they also come with many security risks. If forged images g...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
POSTS&TELECOM PRESS Co., LTD
2024-12-01
|
| Series: | 网络与信息安全学报 |
| Subjects: | |
| Online Access: | http://www.cjnis.com.cn/thesisDetails#10.11959/j.issn.2096-109x.2024083 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1823864847010365440 |
|---|---|
| author | HUANG Jinkun HUANG Yuanhang HUANG Wenmin LUO Weiqi |
| author_facet | HUANG Jinkun HUANG Yuanhang HUANG Wenmin LUO Weiqi |
| author_sort | HUANG Jinkun |
| collection | DOAJ |
| description | With the maturation of computer graphics (CG) technology in the field of image generation, the realism of created images has been improved significantly. Although these technologies are widely used in daily life and bring many conveniences, they also come with many security risks. If forged images generated using CG technology are maliciously used and widely spread on the Internet and social media, they may harm the rights of individuals and enterprises. Therefore, an innovative cross-stream attention enhanced central difference convolutional network was proposed, aiming at improving the accuracy of CG image detection. A dual-stream structure was constructed in the model, in order to extract semantic features and non-semantic residual texture features from the image. Vanilla convolutional layers in each stream were replaced by central difference convolutions, which allowed the model to simultaneously extract pixel intensity information and pixel gradient information from the image. Furthermore, by introducing a cross-stream attention enhancement module, the model enhanced feature extraction capability at the global level and promoted complementarity between the two feature streams. Experimental results demonstrate that this method outperforms existing methods. Additionally, a series of ablation experiments further verify the rationality of the proposed model design. |
| format | Article |
| id | doaj-art-467e3689a7de42888587c120e2c1eaaa |
| institution | Kabale University |
| issn | 2096-109X |
| language | English |
| publishDate | 2024-12-01 |
| publisher | POSTS&TELECOM PRESS Co., LTD |
| record_format | Article |
| series | 网络与信息安全学报 |
| spelling | doaj-art-467e3689a7de42888587c120e2c1eaaa2025-02-08T19:00:08ZengPOSTS&TELECOM PRESS Co., LTD网络与信息安全学报2096-109X2024-12-01109610880361643Cross-stream attention enhanced central difference convolutional network for CG image detectionHUANG JinkunHUANG YuanhangHUANG WenminLUO WeiqiWith the maturation of computer graphics (CG) technology in the field of image generation, the realism of created images has been improved significantly. Although these technologies are widely used in daily life and bring many conveniences, they also come with many security risks. If forged images generated using CG technology are maliciously used and widely spread on the Internet and social media, they may harm the rights of individuals and enterprises. Therefore, an innovative cross-stream attention enhanced central difference convolutional network was proposed, aiming at improving the accuracy of CG image detection. A dual-stream structure was constructed in the model, in order to extract semantic features and non-semantic residual texture features from the image. Vanilla convolutional layers in each stream were replaced by central difference convolutions, which allowed the model to simultaneously extract pixel intensity information and pixel gradient information from the image. Furthermore, by introducing a cross-stream attention enhancement module, the model enhanced feature extraction capability at the global level and promoted complementarity between the two feature streams. Experimental results demonstrate that this method outperforms existing methods. Additionally, a series of ablation experiments further verify the rationality of the proposed model design.http://www.cjnis.com.cn/thesisDetails#10.11959/j.issn.2096-109x.2024083computer graphicsCG image detectioncentral difference convolutionattention mechanism |
| spellingShingle | HUANG Jinkun HUANG Yuanhang HUANG Wenmin LUO Weiqi Cross-stream attention enhanced central difference convolutional network for CG image detection 网络与信息安全学报 computer graphics CG image detection central difference convolution attention mechanism |
| title | Cross-stream attention enhanced central difference convolutional network for CG image detection |
| title_full | Cross-stream attention enhanced central difference convolutional network for CG image detection |
| title_fullStr | Cross-stream attention enhanced central difference convolutional network for CG image detection |
| title_full_unstemmed | Cross-stream attention enhanced central difference convolutional network for CG image detection |
| title_short | Cross-stream attention enhanced central difference convolutional network for CG image detection |
| title_sort | cross stream attention enhanced central difference convolutional network for cg image detection |
| topic | computer graphics CG image detection central difference convolution attention mechanism |
| url | http://www.cjnis.com.cn/thesisDetails#10.11959/j.issn.2096-109x.2024083 |
| work_keys_str_mv | AT huangjinkun crossstreamattentionenhancedcentraldifferenceconvolutionalnetworkforcgimagedetection AT huangyuanhang crossstreamattentionenhancedcentraldifferenceconvolutionalnetworkforcgimagedetection AT huangwenmin crossstreamattentionenhancedcentraldifferenceconvolutionalnetworkforcgimagedetection AT luoweiqi crossstreamattentionenhancedcentraldifferenceconvolutionalnetworkforcgimagedetection |