Advancements in Face Recognition and Deepfake Detection

The field of face recognition and deepfake detection is rapidly evolving, with a focus on improving the accuracy and robustness of models. Recent developments have highlighted the importance of considering recognizability, a key perceptual factor in human face processing, to enhance feature representation and improve the performance of face recognition systems. Additionally, there is a growing need for explainable and transparent models, particularly in the context of deepfake detection, where the ability to provide verifiable reasoning explanations is crucial. Researchers are also exploring new approaches to detect and mitigate adversarial attacks, including the use of multimodal large language models and ensemble-based methods. Noteworthy papers in this area include QCFace, which introduces a hard margin strategy to improve face recognition, and EDVD-LLaMA, which proposes an explainable deepfake video detection framework. Furthermore, papers like Latent Feature Alignment and Fake-in-Facext have made significant contributions to the discovery of biased subpopulations in face recognition models and fine-grained explainable deepfake analysis, respectively.

Sources

QCFace: Image Quality Control for boosting Face Representation & Recognition

Latent Feature Alignment: Discovering Biased and Interpretable Subpopulations in Face Recognition Models

Unmasking Facial DeepFakes: A Robust Multiview Detection Framework for Natural Images

EDVD-LLaMA: Explainable Deepfake Video Detection via Multimodal Large Language Model Reasoning

Investigating Adversarial Robustness against Preprocessing used in Blackbox Face Recognition

Optimizing DINOv2 with Registers for Face Anti-Spoofing

iDETEX: Empowering MLLMs for Intelligent DETailed EXplainable IQA

Fair and Interpretable Deepfake Detection in Videos

Can You Trust What You See? Alpha Channel No-Box Attacks on Video Object Detection

Explainable Face Presentation Attack Detection via Ensemble-CAM

A new wave of vehicle insurance fraud fueled by generative AI

Reliable and Reproducible Demographic Inference for Fairness in Face Analysis

Fake-in-Facext: Towards Fine-Grained Explainable DeepFake Analysis

Built with on top of