A Dense U-Net with Cross-Layer Intersection for Detection and Localization of Image Forgery

In this paper, we apply cross-layer intersection mechanism to dense u-net for image forgery detection and localization. We first train DenseNet for binary classification. Spatial rich model (SRM) filters are adopted for capturing residual signals in the detected images. Then we propose a new approach to preserve complete feature maps of fully connected layer and consider them as the spatial decision information for image segmentation. In addition, these features in downsampling path are transferred more effectively and densely to upsampling path through multiscale upsampling and concatenation. A multi-stage training scheme is then applied to improve the convergence of the network. The experimental results show that the proposed method works well on several standard datasets.

[1]  Kilian Q. Weinberger,et al.  Densely Connected Convolutional Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[2]  Babak Mahdian,et al.  Using noise inconsistencies for blind image forensics , 2009, Image Vis. Comput..

[3]  Muhammad Ghulam,et al.  Image forgery detection using steerable pyramid transform and local binary pattern , 2013, Machine Vision and Applications.

[4]  Jiwu Huang,et al.  A survey of passive technology for digital image forensics , 2007, Frontiers of Computer Science in China.

[5]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[6]  Shih-Fu Chang,et al.  Detecting Image Splicing using Geometry Invariants and Camera Characteristics Consistency , 2006, 2006 IEEE International Conference on Multimedia and Expo.

[7]  Pan Feng,et al.  A survey of passive technology for digital image forensics , 2007 .

[8]  Jessica J. Fridrich,et al.  Rich Models for Steganalysis of Digital Images , 2012, IEEE Transactions on Information Forensics and Security.

[9]  B. S. Manjunath,et al.  Detection and Localization of Image Forgeries Using Resampling Features and Deep Learning , 2017, 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[10]  Zulfiqar Habib,et al.  Copy-move and splicing image forgery detection and localization techniques: a review , 2017 .

[11]  Davide Cozzolino,et al.  Splicebuster: A new blind image splicing detector , 2015, 2015 IEEE International Workshop on Information Forensics and Security (WIFS).

[12]  Ee-Chien Chang,et al.  Detecting Digital Image Forgeries by Measuring Inconsistencies of Blocking Artifact , 2007, 2007 IEEE International Conference on Multimedia and Expo.

[13]  Nasir D. Memon,et al.  Image tamper detection based on demosaicing artifacts , 2009, 2009 16th IEEE International Conference on Image Processing (ICIP).

[14]  Jiangqun Ni,et al.  A deep learning approach to detection of splicing and copy-move forgeries in images , 2016, 2016 IEEE International Workshop on Information Forensics and Security (WIFS).

[15]  Larry S. Davis,et al.  Learning Rich Features for Image Manipulation Detection , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[16]  C.-C. Jay Kuo,et al.  Image Splicing Localization using a Multi-task Fully Convolutional Network (MFCN) , 2017, J. Vis. Commun. Image Represent..

[17]  Longin Jan Latecki,et al.  Dense Deconvolutional Network for Semantic Segmentation , 2018, 2018 25th IEEE International Conference on Image Processing (ICIP).

[18]  Trevor Darrell,et al.  Fully Convolutional Networks for Semantic Segmentation , 2017, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[19]  Jing Dong,et al.  CASIA Image Tampering Detection Evaluation Database , 2013, 2013 IEEE China Summit and International Conference on Signal and Information Processing.

[20]  Ying Zhang,et al.  Image Region Forgery Detection: A Deep Learning Approach , 2016, SG-CRC.