Abstract:Image forgery detection aims to detect and locate forged regions in an image. Most existing forgery detection algorithms formulate classification problems to classify pixels into forged or pristine. However, the definition of forged and pristine pixels is only relative within one single image, e.g., a forged region in image A is actually a pristine one in its source image B (splicing forgery). Such a relative definition has been severely overlooked by existing methods, which unnecessarily mix forged (pristine) regions across different images into the same category. To resolve this dilemma, we propose the FOrensic ContrAstive cLustering (FOCAL) method, a novel, simple yet very effective paradigm based on contrastive learning and unsupervised clustering for the image forgery detection. Specifically, FOCAL 1) utilizes pixel-level contrastive learning to supervise the high-level forensic feature extraction in an image-by-image manner, explicitly reflecting the above relative definition; 2) employs an on-the-fly unsupervised clustering algorithm (instead of a trained one) to cluster the learned features into forged/pristine categories, further suppressing the cross-image influence from training data; and 3) allows to further boost the detection performance via simple feature-level concatenation without the need of retraining. Extensive experimental results over six public testing datasets demonstrate that our proposed FOCAL significantly outperforms the state-of-the-art competing algorithms by big margins: +24.3% on Coverage, +18.6% on Columbia, +17.5% on FF++, +14.2% on MISD, +13.5% on CASIA and +10.3% on NIST in terms of IoU. The paradigm of FOCAL could bring fresh insights and serve as a novel benchmark for the image forgery detection task. The code is available at <a class="link-external link-https" href="https://github.com/HighwayWu/FOCAL" rel="external noopener nofollow">this https URL</a>.

Detecting Image Forgery Using Perspective Constraints

Document Forgery Detection Using Distortion Mutation of Geometric Parameters in Characters

Digital Image Forgery Detection Based on Lens and Sensor Aberration

Inter-frame Passive-Blind Forgery Detection for Video Shot Based on Similarity Analysis

Spotting the Difference: Context Retrieval and Analysis for Improved Forgery Detection and Localization

Revisiting copy-move forgery detection by considering realistic image with similar but genuine objects

Image Forgery Detection: Survey and Future Directions

Anti-Compression Contrastive Facial Forgery Detection

A Robust Approach to Detect Digital Forgeries by Exploring Correlation Patterns

Estimation of Image Rotation Angle Using Interpolation-Related Spectral Signatures with Application to Blind Detection of Image Forgery

Detecting Image Forgeries using Geometric Cues

Rethinking Image Forgery Detection via Contrastive Learning and Unsupervised Clustering

Detection of image recognition forgery technology under machine vision

Survey on Blind Image Forgery Detection

An Evaluation of Digital Image Forgery Detection Approaches

Copy-move forgery detection for image forensics using the superpixel segmentation and the Helmert transformation

Discrepancy-Guided Reconstruction Learning for Image Forgery Detection

Fogery image blind detection by asymemetric search based on SIFT

Low-complexity Fake Face Detection Based on Forensic Similarity

Boundary-based Image Forgery Detection by Fast Shallow CNN

Exploring Disentangled Content Information for Face Forgery Detection