Audio Deepfake Detection: ICASSP 2022

Associated SPS Event: IEEE ICASSP 2022 Grand Challenge

Over the last few years, the technology of speech synthesis and voice conversion has made significant improvement with the development of deep learning. The models can generate realistic and human-like speech. It is difficult for most people to distinguish the generated audio from the real. However, this technology also poses a great threat to the global political economy and social stability if some attackers and criminals misuse it with the intent to cause harm. The ASVspoof challenges have played a key role in fostering spoofed speech detection research for protect automatic speaker verification systems from manipulation. Although the audio deep synthesis detection task is included in the ASVspoof 2021, it mainly involves compressed audio similar to the LA task However, it ignores some challenging attacking situations in realistic scenarios: (1) diverse background noises and disturbances are contained in the fake audios. (2) several small fake clips are hided in a real speech audio. (3) new algorithms of speech synthesis and voice conversion are proposed rapidly. These pose a serious threat since that it is difficult to deal with the above-mentioned attacking situations. Therefore, we launch the first Audio Deep Synthesis Detection Challenge (ADD 2022) to spur researchers around the world to build innovative new technologies that can further accelerate and foster research on detecting deep synthesis and manipulated audios. The ADD 2022 includes three tracks. The ADD 2022 is an IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2022 Signal Processing Grand Challenge. Five papers from the top-ranked teams are selected to be submitted via the ICASSP paper submission system for review before final acceptance. The format should be consistent with ICASSP 2022 regular paper. Top three winning teams from each track will be awarded as outlined in the description of the rules.