The field of deepfake detection and image forgery localization is rapidly advancing, with a focus on developing robust and generalizable methods to combat the increasing sophistication of image manipulation techniques. Researchers are exploring innovative approaches, including the use of hybrid CNN-Transformer models, Vision Transformers, and multimodal large language models, to improve detection accuracy and localization precision. Noteworthy papers in this area include EdgeDoc, which presents a novel approach for detecting and localizing document forgeries, and Veritas, which introduces a multi-modal large language model-based deepfake detector with pattern-aware reasoning. Additionally, papers such as A Novel Local Focusing Mechanism for Deepfake Detection Generalization and No Pixel Left Behind: A Detail-Preserving Architecture for Robust High-Resolution AI-Generated Image Detection are making significant contributions to the field by proposing new mechanisms and architectures for improving detection performance.
Deepfake Detection and Image Forgery Localization
Sources
EdgeDoc: Hybrid CNN-Transformer Model for Accurate Forgery Detection and Localization in ID Documents
Seeing Isn't Believing: Addressing the Societal Impact of Deepfakes in Low-Tech Environments
Combating Digitally Altered Images: Deepfake Detection
A Novel Local Focusing Mechanism for Deepfake Detection Generalization
No Pixel Left Behind: A Detail-Preserving Architecture for Robust High-Resolution AI-Generated Image Detection
Edge-Enhanced Vision Transformer Framework for Accurate AI-Generated Image Detection
Propose and Rectify: A Forensics-Driven MLLM Framework for Image Manipulation Localization
Addressing Deepfake Issue in Selfie banking through camera based authentication
Improving Generalization in Deepfake Detection with Face Foundation Models and Metric Learning
SDiFL: Stable Diffusion-Driven Framework for Image Forgery Localization
Webly-Supervised Image Manipulation Localization via Category-Aware Auto-Annotation
Veritas: Generalizable Deepfake Detection via Pattern-Aware Reasoning
FakeParts: a New Family of AI-Generated DeepFakes