A contemporary survey on deepfake detection: datasets, algorithms, and challenges. Liang Yu Gong and Xue Jun Li review the state-of-the-art methods for detecting deepfakes, focusing on facial forgery detection from 2019 to 2023. They categorize detection methods into four groups: traditional CNN-based detection, CNN backbone with semi-supervised detection, transformer-based detection, and biological signal detection. The authors summarize several representative deepfake detection datasets, including FaceForensics++, DFDC, and Celeb-DF V2, and evaluate the performance of detection models across these datasets. They find that cross-dataset evaluation significantly degrades accuracy, highlighting the importance of robustness in detection models. The survey also discusses recent trends in deepfake detection, such as biological signal-based detection, and presents a comprehensive analysis of various detection methods, including Capsule Network, CORE, and T-Face. The authors emphasize the need for further research to develop more reliable detection algorithms. The survey provides a detailed comparison of evaluation metrics, including accuracy, AUC, and EER, and highlights the challenges in deepfake detection, such as the difficulty in distinguishing between real and fake videos. The study concludes that deepfake detection is a rapidly evolving field with significant challenges, and further research is needed to improve detection accuracy and robustness.A contemporary survey on deepfake detection: datasets, algorithms, and challenges. Liang Yu Gong and Xue Jun Li review the state-of-the-art methods for detecting deepfakes, focusing on facial forgery detection from 2019 to 2023. They categorize detection methods into four groups: traditional CNN-based detection, CNN backbone with semi-supervised detection, transformer-based detection, and biological signal detection. The authors summarize several representative deepfake detection datasets, including FaceForensics++, DFDC, and Celeb-DF V2, and evaluate the performance of detection models across these datasets. They find that cross-dataset evaluation significantly degrades accuracy, highlighting the importance of robustness in detection models. The survey also discusses recent trends in deepfake detection, such as biological signal-based detection, and presents a comprehensive analysis of various detection methods, including Capsule Network, CORE, and T-Face. The authors emphasize the need for further research to develop more reliable detection algorithms. The survey provides a detailed comparison of evaluation metrics, including accuracy, AUC, and EER, and highlights the challenges in deepfake detection, such as the difficulty in distinguishing between real and fake videos. The study concludes that deepfake detection is a rapidly evolving field with significant challenges, and further research is needed to improve detection accuracy and robustness.