Detect and Locate : Exposing Face Manipulation by Semantic- and Noise-level Telltales

Chenqi KONG, Baoliang CHEN, Haoliang LI, Shiqi WANG, Anderson ROCHA, Sam KWONG

Research output: Journal PublicationsJournal Article (refereed)peer-review

31 Citations (Scopus)


The technological advancements of deep learning have enabled sophisticated face manipulation schemes, raising severe trust issues and security concerns in modern society. Generally speaking, detecting manipulated faces and locating the potentially altered regions are challenging tasks. Herein, we propose a conceptually simple but effective method to efficiently detect forged faces in an image while simultaneously locating the manipulated regions. The proposed scheme relies on a segmentation map that delivers meaningful high-level semantic information clues about the image. Furthermore, a noise map is estimated, playing a complementary role in capturing low-level clues and subsequently empowering decision-making. Finally, the features from these two modules are combined to distinguish fake faces. Extensive experiments show that the proposed model achieves state-of-the-art detection accuracy and remarkable localization performance.
Original languageEnglish
Pages (from-to)1741-1756
JournalIEEE Transactions on Information Forensics and Security
Early online date28 Apr 2022
Publication statusPublished - 2022
Externally publishedYes


  • Face forensics
  • face forgery detection
  • face manipulation localization


Dive into the research topics of 'Detect and Locate : Exposing Face Manipulation by Semantic- and Noise-level Telltales'. Together they form a unique fingerprint.

Cite this