Get our free extension to see links to code for papers anywhere online!

 Add to Chrome

 Add to Firefox

CatalyzeX Code Finder - Browser extension linking code for ML papers across the web! | Product Hunt Embed

D-Unet: A Dual-encoder U-Net for Image Splicing Forgery Detection and Localization

Dec 03, 2020
Xiuli Bi, Yanbin Liu, Bin Xiao, Weisheng Li, Chi-Man Pun, Guoyin Wang, Xinbo Gao

Recently, many detection methods based on convolutional neural networks (CNNs) have been proposed for image splicing forgery detection. Most of these detection methods focus on the local patches or local objects. In fact, image splicing forgery detection is a global binary classification task that distinguishes the tampered and non-tampered regions by image fingerprints. However, some specific image contents are hardly retained by CNN-based detection networks, but if included, would improve the detection accuracy of the networks. To resolve these issues, we propose a novel network called dual-encoder U-Net (D-Unet) for image splicing forgery detection, which employs an unfixed encoder and a fixed encoder. The unfixed encoder autonomously learns the image fingerprints that differentiate between the tampered and non-tampered regions, whereas the fixed encoder intentionally provides the direction information that assists the learning and detection of the network. This dual-encoder is followed by a spatial pyramid global-feature extraction module that expands the global insight of D-Unet for classifying the tampered and non-tampered regions more accurately. In an experimental comparison study of D-Unet and state-of-the-art methods, D-Unet outperformed the other methods in image-level and pixel-level detection, without requiring pre-training or training on a large number of forgery images. Moreover, it was stably robust to different attacks.

* 13 pages, 13 figures 

Share this with someone who'll enjoy it:

   Access Paper Source

Share this with someone who'll enjoy it: