Zhiwu Huang

There have been emerging a number of benchmarks and techniques for the detection of deepfakes. However, very few works study the detection of incrementally appearing deepfakes in the real-world scenarios. To simulate the wild scenes, this paper suggests a continual deepfake detection benchmark (CDDB) over a new collection of deepfakes from both known and unknown generative models. The suggested CDDB designs multiple evaluations on the detection over easy, hard, and long sequence of deepfake tasks, with a set of appropriate measures. In addition, we exploit multiple approaches to adapt multiclass incremental learning methods, commonly used in the continual visual recognition, to the continual deepfake detection problem. We evaluate existing methods, including their adapted ones, on the proposed CDDB. Within the proposed benchmark, we explore some commonly known essentials of standard continual learning. Our study provides new insights on these essentials in the context of continual deepfake detection. The suggested CDDB is clearly more challenging than the existing benchmarks, which thus offers a suitable evaluation avenue to the future research. Our benchmark dataset and the source code will be made publicly available.

A Continual Deepfake Detection Benchmark: Dataset, Methods, and Essentials. Chuqiao Li, Zhiwu Huang*, Danda Pani Paudel , Yabin Wang, Mohamad Shahbazi, Xiaopeng Hong, Luc Van Gool . (*indicates corresponding author) . In Winter Conference on Applications of Computer Vision (WACV), 2023. Preprint

We use Hollywood movie trailers for the dataset collection, which was motivated by the fact movie trailers highlight dramatic and emotionally charged scenes. Unlike whole movies, interviews or TV series, trailers contain scenes where stronger emotional response of actors are highlighted. Furthermore, using trailers of thousands of movies increases the gender, racial and age-wise diversity of the faces in the clips. Approximately 6000 complete Hollywood movie trailers were downloaded from YouTube. Number of SIFT feature matches between corresponding frames was used for shot boundary detection. Approximately 200, 000 shots were detected in those trailers. Concretly, the dataset contains approximately 200,000 individual clips of various facial expressions, where the faces are cropped with 256x256 resolution from about 6,000 high resolution movie trailers on YouTube. We convert them to tfrecord with resolutions range from 4x4 to 256x256.

Towards High Resolution Video Generation with Progressive Growing of Sliced Wasserstein GANs. Dinesh Acharya, Zhiwu Huang, Danda Pani Paudel , Luc Van Gool . arXiv preprint arXiv:1810.02419, 2018. Paper