TSG: Time Step Generating
We proposed a general-purpose deepfake image detector called Time Step Generating (TSG). Unlike existing methods, TSG is independent of pretraining model reconstruction abilities, specific datasets, or sampling algorithms.
We further extended this deepfake detection technique to fake video detection, enabling localization of manipulated frames and forged regions.
Published: ACM MMAsia 2025 (Oral)
ZIYUE ZENG
Master’s Student → Incoming PhD Student at Waseda University
Pursuing socially meaningful, grounded research with enduring passion and diligence.