Exploiting Spatial Structure for Localizing Manipulated Image Regions
read more
Citations
FaceForensics++: Learning to Detect Manipulated Facial Images
FaceForensics++: Learning to Detect Manipulated Facial Images
Face X-Ray for More General Face Forgery Detection
Multi-task Learning for Detecting and Segmenting Manipulated Facial Images and Videos
Learning Rich Features for Image Manipulation Detection
References
Adam: A Method for Stochastic Optimization
Fully convolutional networks for semantic segmentation
Fast R-CNN
Fast R-CNN
SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation
Related Papers (5)
A Deep Learning Approach to Universal Image Manipulation Detection Using a New Convolutional Layer
A deep learning approach to detection of splicing and copy-move forgeries in images
Frequently Asked Questions (14)
Q2. What is the key insight of using LSTM?
The key insight of using LSTM is to learn the boundary transformation between different blocks, which provides discriminative features between manipulated and non-manipulated regions.
Q3. What is the common method used in image forensics?
In image forensics, most of the state-of-the-art image tamper detection approaches exploit the frequency domain characteristics and/or statistical properties of an image.
Q4. What is the main reason why deep learning is popular?
deep learning has become popular due to its promising performance in different visual recognition tasks such as object detection [26, 8], scene classification [62], and semantic segmentation [40].
Q5. How do the authors learn the patch tasks?
The authors perform end-to-end training to learn thejoint tasks through back-propagation using ground-truth patch labels and mask information.
Q6. What is the method for minimizing the loss of the network?
The authors use adaptive moment estimation (Adam) [32] optimization technique in order to minimize the total loss of the network, shown in Eqn.
Q7. What is the main topic of the paper?
The paper [28] includes the recent advances in image manipulation and discusses the process of restoring missing or damaged areas in an image.
Q8. What is the definition of copy-move forgeries?
In order to detect copy-move forgeries, an image is first divided into overlapping blocks and some sort of distance measure or correlation is used to determine blocks that have been cloned.
Q9. What are some types of manipulations that can easily deceive the human perceptual system?
There are certain types of manipulations such as copy-move, splicing, removal, that can easily deceive the human perceptual system.
Q10. What is the purpose of using convolutional layers?
These convolutional layers learn the mapping between features of the boundary transformation from the LSTM and the tampered pixels using the ground-truth mask.
Q11. What is the purpose of this paper?
In this paper, the authors present a unified framework for joint patch classification and segmentation to localize manipulated regions from an image.
Q12. How many patches are used in the training dataset?
In data preparation, the authors first split the whole image dataset into three subsets- training (65%), validation (10%) and testing (25%).
Q13. How many feature maps do the authors use in their proposed network?
The authors also try with varying number of feature maps such as (1) Conv1-8f : conv1 with 8 maps, (2) Conv1-32f : conv1 layer with 32 feature maps.
Q14. What is the novel image forgery detection method?
A novel image forgery detection method is presented in [46] based on the steerable pyramid transform (SPT) and the local binary pattern (LBP).