Abstract:As an ill-posed problem, median filtered (MF) image restoration and anti-forensics still face a challenge in improving the visual quality and deceiving the forensic detectors simultaneously. MF image restoration mainly aims at image visual quality recovery, while MF image anti-forensics primarily focuses on deceiving the existing forensic algorithms. Median filtering is popularly used as an operation for image denoising and smoothing. It is also employed by image anti-forensics to conceal or remove the traces of other image manipulating operations, such as JPEG blocking artifacts and the periodicity introduced by image resampling. This work proposes an adversarial learning framework that introduces loss terms spanning three domains for MF image restoration and anti-forensics. The differences of high-frequency coefficients in the discrete wavelet transform (DWT) domain are introduced as a term of loss term to restore high-frequency components. Moreover, high-frequency subband variance loss in the discrete cosine transform (DCT) domain is proposed to make the statistical characteristics of the generated image closer to that of the original image. Together with the Huber loss in the spatial domain, a novel generator network is developed and trained to enhance information flow and feature reuse. A capsule network is applied as a discriminator for differentiating the generated image from the original version. We explore a solution through adversarial learning in an infinite set of possible solutions. Experimental results demonstrate that the proposed method is able to restore the visual quality of MF images with salt and pepper noise previously more effectively than the existing methods. What is more, our method achieves the best anti-forensic performance in comparison with the existing state-of-the-art MF image anti-forensic methods, while enhancing the quality of anti-forensically restored images.

Robust Median Filtering Forensics by CNN-Based Multiple Residuals Learning

Median Filtering Forensics Based on Convolutional Neural Networks

Frequency-Relevant Residual Learning for Multi-Modal Image Denoising.

An end-to-end deep learning model for robust smooth filtering identification

An adversarial learning framework with cross-domain loss for median filtered image restoration and anti-forensics

Image Denoising Via Multi-Scale Gated Fusion Network

Blind Forensics of Median Filtering Based on Markov Statistics in Median-Filtered Residual Domain

A multi-purpose image forensic method using densely connected convolutional neural networks

CNN based image resizing forensics for double compressed JPEG images

Training CNNs in Presence of JPEG Compression: Multimedia Forensics vs Computer Vision

Median Pixel Difference Convolutional Network for Robust Face Recognition

Deep Multi-scale Discriminative Networks for Double JPEG Compression Forensics

Image Forgery Detection and Localization via a Reliability Fusion Map

Robust Contrast Enhancement Forensics Using Pixel and Histogram Domain CNNs

Accurate and lightweight dehazing via multi-receptive-field non-local network and novel contrastive regularization

Image Denoising Based On A Cnn Model

Dmcnn: Dual-Domain Multi-Scale Convolutional Neural Network for Compression Artifacts Removal

Refining Localized Attention Features with Multi-Scale Relationships for Enhanced Deepfake Detection in Spatial-Frequency Domain

A Convolutional Neural Network-Based Conditional Random Field Model for Structured Multi-Focus Image Fusion Robust to Noise

Image Processing Operations Identification via Convolutional Neural Network

Multiple Manipulation Detection in Images Using Frequency Domain Features in 3D-CNN