A Multi-purpose Image Counter-anti-forensic Method Using Convolutional Neural Networks

Interpol review of imaging and video 2016–2019

“…The main difference between image recognition and image manipulation detection is the signal strength. Image manipulation detection, in contrast to image recognition, has to cope with very small differences between the manipulated image and the original image [ 13 ].…”

“…Yu et al. [ 13 ], developed a 5-layer regular CNN model consisting of four convolutional layers, the second and fourth convolutional layer followed by a max-pooling layers, and one fully connected layer with softmax activation function. There model was trained and tested to detect anti-JPEG compression [ 26 , 27 ], anti-median filtering [ 23 , 24 ], antiresampling [ 25 ] and anti-contrast enhancement [ 22 ].…”

“…According to Yu et al. [ 13 ] filter size of the convolutional layer is critical for the network performance. The more suitable the receptive filter the better the resulting extracted features.…”

Bibliography of digital image anti‐forensics and anti‐anti‐forensics techniques

“…Unlike traditional approaches which are based on hand-crafted features which are laborious and require domain knowledge, deep learning can learn to extract important features automatically from raw images. Yu et al [147] proposed a framework using a CNN to counter multiple anti-forensics. Based on the automatic feature extraction, the approach was capable of detecting various kind of anti-forensics attacks in a tampered image and showed superior performance compared to well-known image AFTs.…”

A Survey of Image Information Hiding Algorithms Based on Deep Learning

“…Reinforcement learning [Mnih, Kavukcuoglu, Silver et al (2015)] emphasizes how to act on the environment to maximize the expected benefits. In application, deep learning has been greatly developed in the fields of video [Feichtenhofer, Fan, Malik et al (2018); Wichers, Villegas, Erhan et al (2018); Wang, Liu, Zhu et al (2018)], image [Xie, He, Zhang et al (2018); Barz and Denzler (2018); Wang and Chan (2018)], voice [Yang, Lalitha, Lee et al (2018); Arik, Chen, Peng et al (2018); Qian, Du, Hou et al (2017)], semantic understanding [Qin, Kamnitsas, Ancha et al (2018); Zhuang and Yang (2018); Sanh, Wolf and Ruder (2018)], and has been further applied in object detection [Roddick, Kendall and Cipolla (2018) ;Jaeger, Kohl, Bickelhaupt et al (2018)], image forensics [Yu, Zhan and Yang (2016), Cui, McIntosh and Sun (2018)], intelligent management [Liang, Jiang, Chen et al (2018); Le, Pham, Sahoo et al (2018); Duan, Lou, Wang et al (2017)] and medicine [Mobadersany, Yousefi, Amgad et al (2018); Rajpurkar, Irvin, Zhu et al (2017); Akkus, Galimzianova, Hoogi et al (2017)]. In the field of supervised learning, image classification methods that based on deep learning have been mature, which can be applied to object detection and image retrieval.…”