https://scholars.lib.ntu.edu.tw/handle/123456789/607199
標題: | Polling Mechanism for Video Deepfake | 作者: | Hsu H.-W Huang C.-W. JIAN-JIUN DING |
關鍵字: | forensic image processing;head orientation;motion;polling mechanism;video deepfake;Digital forensics;Accuracy rate;Forensic image processing;Frame differences;Frame-based;Head orientation;Images processing;Learning-based methods;Motion;Polling mechanism;Video deepfake;Video signal processing | 公開日期: | 2021 | 起(迄)頁: | 320-323 | 來源出版物: | Proceedings of the 3rd IEEE Eurasia Conference on IOT, Communication and Engineering 2021, ECICE 2021 | 摘要: | A polling-based mechanism is developed to identify whether a video is forged. It is important for forensic image processing. However, most of the existing video deepfake algorithms are frame-based. In other words, a learning-based method is applied to identify whether a frame is forged then the mean of the fake score for all frames is applied to determine whether the whole video is forged. In this work, we propose a polling mechanism to well integrate the deepfake score of each frame. We found that the misidentification of a deepfake algorithm usually occurs in the frame with drastic motion, a tilted head, and blinking eyes. Therefore, we determine the weights of each frame according to the frame difference, the head orientation, whether the eyes are blinking, and the accuracy rate of the validation data. With the proposed polling mechanism, the accuracy of video deepfake can be improved and whether a video is forged can be well determined using a much smaller number of frames. ? 2021 IEEE. |
URI: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85124193980&doi=10.1109%2fECICE52819.2021.9645661&partnerID=40&md5=623243e8a12d23d66598818382ab0c5c https://scholars.lib.ntu.edu.tw/handle/123456789/607199 |
DOI: | 10.1109/ECICE52819.2021.9645661 |
顯示於: | 電機工程學系 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。