חזרה לעמוד הקודם

TAR: Generalized Forensic Framework to Detect Deepfakes using Weakly Supervised Learning (2021)

Sangyup Lee, Shahroz Tariq, Junyaup Kim, and Simon S. Woo, Sungkyunkwan University, Suwon, South Korea (2021)

מאמר עדכני למאי 2021 מטעם אוניברסיטת סנגקיונקואן בקוריאה הדרומית, אודות מערכת לזיהוי חומרי DEEPFAKE המיועדת לקודד את החומרים שניתנים לה בצורה אוטומטית (Transfer learning-based Auto encoder with Residuals- TAR). המחקר התבסס על המאגרים "FaceForensics++" ו-"Deepfake-in-the-Wild" ושואף לפתח את המערכת לנקודה בה תוכל לזהות חומרי DEEPFAKE באיכות גבוהה במיוחד תוך התבססות על מספר נמוך יחסית של חומרי DEEPFAKE שיוזנו לתוכה מראש.

abstract:
Deepfakes have become a critical social problem, and detecting them is of utmost importance. Also, deepfake generation methods are advancing, and it is becoming harder to detect. While many deepfake
detection models can detect different types of deepfakes separately, they perform poorly on generalizing the detection performance over multiple types of deepfake. This motivates us to develop a generalized model to detect different types of deepfakes. Therefore, in this work, we introduce a practical digital forensic tool to detect different types of deepfakes simultaneously and propose Transfer learning-based Autoencoder with Residuals (TAR). The ultimate goal of our work is to develop a unified model to detect various types of deepfake videos with high accuracy, with only a small number of training samples that can work well in realworld settings. We develop an autoencoder-based detection model with Residual blocks and sequentially perform transfer learning to detect different types of deepfakes simultaneously. Our approach achieves a much higher generalized detection performance than the state-of-the-art methods on the FaceForensics++ dataset. In addition, we evaluate our model on 200 real-world Deepfake-in-the-Wild (DW) videos of 50 celebrities available on the Internet and achieve 89.49% zero-shot accuracy, which is significantly higher than the best baseline model (gaining 10.77%),
demonstrating and validating the practicability of our approach.

קישור למאמר המלא