siwei lyu image

Siwei Lyu, Ph.D. [CV] [Google Scholar] [DBLP]

Fellow of IEEE and IAPR
SUNY Empire Innovation Professor, Department of Computer Science and Engineering
Director, UB Media Forensic Lab (UB MDFL)
Co-Director, Center for Information Integrity (CII)
School of Engineering and Applied Sciences
University at Buffalo, State University of New York
Phone: (716) 437-1587
Email: siweilyu AT buffalo DOT edu
Office: 317 Davis Hall, Buffalo, NY14260

  • (08/06/2021) I gave a keynote speech at the 2021 International Joint Conference on Biometrics (IJCB 2021)
  • (07/26/2021) I was featured in an online article on fake face generation [Link].
  • (06/18/2021) I was invited to give a Keynote speech at the CVPR Workshop on Multimedia Forensics on our work of using corneal spectral highlights to detect GAN synthesized faces. [Link to video].
  • (06/08/2021) With my former student/post-doc Yuezun Li, I gave a tutorial at ICASSP 2021 on DeepFake Generation and Detection [Part 1] [Part 2]. [Part 3].
  • (03/15/2021) Our work on detecting GAN generated images using corneal reflections is reported by UB News. [Link]. This work has also been featured on WKBW-TV [Link].
  • (01/29/2021) I appeared on the Interview by Newsy Cable Network and talked about my work on DeepFake forensics. [Link].
  • (12/09/2020) My work on DeepFake detection is listed as one of the "Four AI technologies that could transform the way we live and work" in a Nature article. [Link].
  • (11/18/2020) I was interviewed by Fox News Cleveland on my work of DeepFake detection. [Link].
  • (11/01/2020) Our open-platform for DeepFake detection, DeepFake-o-meter is online now. [Link].
  • (08/11/2020) I was interviewed by BYU Radio "The Lisa Show" to discuss DeepFakes [Link].
  • (07/20/2020) I was featured in a Scientific American Program on DeepFakes [Link], accompanied by a SciAm blog piece written by me [Link].
  • (06/15/2020) Yuezun Li successfully defended his thesis, Congratulations Dr. Li! Yuezun is the sixth Ph.D. student I graduated, along with Xunyu Pan (2011), Amanda Fernandez (2013), Yueming Yang (2014), Xin Wang (2015), and Wenbo Li (2019).
  • (05/26/2020) I was interviewed by Japanese TV Channel NHK's Koukusai Houdou (国際報道) for our work on DeepFake detection [Link].
  • (02/24/2020) Our paper on the [Celeb-DF dataset], which includes 5000+ high-quality synthetic videos has been accepted by CVPR 2020!
  • (02/24/2020) I was featured in Episode 6 of the PBS Documentary "Fake: Searching for Truth in the Age of Misinformation" [Link].
  • (01/30/2020) I was interviewed by Inverse [Link].
  • (01/28/2020) I presented and spoke as a panelist at the FTC workshop on voice cloning technologies [Link] and corresponding report on VentureBeat [Link].
  • (01/08/2020) I was interviewed Fox News about Facebooks's new policy on DeepFakes [Link].
  • (11/26/2019) My comments were featured in an article in the Wired Magazine UK [Link].
  • (11/22/2019) I testified at New York State Senate Hearing on protecting consumer data and privacy on online platforms [Link].
  • (10/09/2019) I was featured in an IEEE Spectrum article [Link].
  • (10/04/2019) I discussed DeepFakes and their detections on WHYY The Pulse [Link].
  • (09/26/2019) I testified at the US House of Representative Committee on Science, Space, and Technology Subcommittee on Oversights and Investigation Hearing on "Online Imposter and Disinformation" Link. See also report on Times Union [Link] and Washington Post [Link].
  • (09/03/2019) I am very excited to be an academic advisor of the DeepFake Detection Challenge sponsored by FaceBook, Microsoft and PAI with multiple universities, along with Profs Hany Farid and Alyosha Efros (UC-Berkeley), Serge Belongie (Cornell Tech), Rama Chellappa (John Hopkins), Phil Torr (Oxford), Phillip Isola and Antonio Torraalba (MIT) [Link]. See also report on Times Union [Link].
  • (07/25/2019) Wenbo Li successfully defended his thesis on Human Pose Estimation and Activity Recognition. Wenbo will join Samsung Research Lab USA as a Research Scientist. Congratulations Dr. Li!
  • (07/11/2019) I was interviewed by Brooklyn Eagles to discuss our work on detecting DeepFakes [Link to story].
  • (06/25/2019) I have written an article on The Conversation about our work of detecting and protecting from DeepFakes [Link].
  • (06/02/2019) Our work on detecting DeepFakes continues to draw attention and is featured on Full Measure with Sharyl Atkison [Link to video] and on Brazil news media O Globo [Link].
  • (05/13/2019) Our work has been covered by Fox News [Link], USA Today [Link] and Chinese National TV (CCTV) [Link].
  • (05/02/2019) Two papers on Digital Media Forensics are accepted by Workshop on Information Hiding and Multimedia Security (IHMMSec2019) in Paris.
  • (05/01/2019) Our lab keeps growing: Dr. Feng Ding joins as a post-doc, Ehab AlBadawy joins as a Ph.D. student, and Ao Luo joins as a visiting student.
  • (04/17/2019) I gave a presentation at O'Reilly AI Conference at NYC [link].
  • (03/01/2019) Our work is featured on Spanish News paper El Pais [link].
  • (02/28/2019) Our work on detecting AI-synthesized fake media has won Google Faculty Research Award for the year of 2019.
  • (02/14/2019) Our work is featured on O'Reily's Data Show Podcast [link] and a Brookings Institute blog [link].
  • (02/01/2019) Our work on detecting DeepFake is featured on Univision [link].
  • (11/19/2018) I received SUNY Chancellor's Award for Research and Creative Activities. Here is a photo at the ceremony.

    Chancellor Award 2018

    (left to right) President Havidàn Rodríguez, Dr. Siwei Lyu, and Dean Kim Boyer.

  • (11/06/2018) Our work on detecting DeepFake generated videos using detection of eye blinking are reported by many news media, including The MIT Technology Review [article], CNN Money [article], Spectrum News [TV interview] Wired Magazine [article], UBuffalo Podcast [interview], Data Skeptic Podcast [interview], Deseret News [article], The Register [article], Columnbia Journalism Review [article], iNews [article]
    • For details, check my article on The Conversation and our paper to be presented at WIFS 18.
  • (09/01/2018) Our pose estimation work is ranked #1 on the very challenging MPII benchmark.
    Video demonstration of our pose estimation results. This work has also been accepted for publication by ECCV.
  • (02/18/2018) I talked about detecting Deep Fakes on Science Friday with Ira Flatlow [interview].
  • (09/06/2017) Paper on Average Top-K Learning is accepted by NIPS 2017.
  • (07/29/2017) I co-organized and participated the IEEE/Nvidia AI City Challenge with the IEEE Smart City Conference 2017. and our team got honary mention.
  • (07/19/2017) Ppaper on using RNN-Tree for Skeleton-based Action Recognition is accepted by ICCV 2017.
  • (07/17/2017) I co-organized the 2017 AVSS Challenge on Advance Traffic Monitoring, and co-organize the International Workshop on Traffic and Street Surveillance and Safety (IWT4S), in conjunction with the 2017 IEEE Conference on Advances in Video and Signal-based Survillance (AVSS).
  • (03/21/2017) I am awarded SUNY Buffalo’s Presidential Award for Excellence in Research and Creative Activities.
  • (02/01/2017) I will be serving as an Area Chair for IEEE International Conference for Image Processing (ICIP) 2017 in Beijing, China.
  • (12/06/2016) I will be serving on IEEE SPS Information Forensics and Security Technical Committee from 2017 to 2019.
  • (10/01/2016) I will be serving as an Associate Editor for IEEE Transactions on Information Forensics and Security from 2016 to 2019.
  • (09/01/2016) I am elevated to IEEE Senior Member.
  • (08/15/2016) Yueming Yang successfully defended his thesis. He also present his work at AVSS 2016 at Colorado Springs. Congratulations, Dr. Yang!