With various facial manipulation techniques arising, face forgery detection has drawn growing attention due to security concerns. Previous works always formulate face forgery detection as a classification problem based on cross-entropy loss, which emphasizes category-level differences rather than the essential discrepancies between real and fake faces, limiting model generalization in unseen domains. To address this issue, we propose a novel face forgery detection framework, named Dual Contrastive Learning (DCL), which specially constructs positive and negative paired data and performs designed contrastive learning at different granularities to learn generalized feature representation. Concretely, combined with the hard sample selection strategy, Inter-Instance Contrastive Learning (Inter-ICL) is first proposed to promote task-related discriminative features learning by especially constructing instance pairs. Moreover, to further explore the essential discrepancies, Intra-Instance Contrastive Learning (Intra-ICL) is introduced to focus on the local content inconsistencies prevalent in the forged faces by constructing local region pairs inside instances. Extensive experiments and visualizations on several datasets demonstrate the generalization of our method against the state-of-the-art competitors. Our Code is available at https://github.com/Tencent/TFace.git.
Density dependent parametrization models of the nucleon-meson effective couplings, including the isovector scalar δ-field, are applied to asymmetric nuclear matter. The nuclear equation of state and the neutron star properties are studied in an effective Lagrangian density approach, using the relativistic mean field hadron theory. It is known that the introduction of a δ-meson in the constant coupling scheme leads to an increase of the symmetry energy at high density and so to larger neutron star masses, in a pure nucleon-lepton scheme. We use here a more microscopic density dependent model of the nucleon-meson couplings to study the properties of neutron star matter and to reexamine the δ-field effects in asymmetric nuclear matter. Our calculations show that, due to the increase of the effective δ coupling at high density, with density dependent couplings the neutron star masses in fact can be even reduced.
One of the most challenging tasks for future high-luminosity electron-positron colliders is to extract Higgs triple coupling. It was proposed that this can be carried out via the precisely measuring the cross section of ZH associated production up to 0.4%. In this paper, we example the possible heavy pollution from Higgs-top anomalous coupling. Our numerical results show that the pollution is small for √ s e + e − = 240GeV . However for the higher energy collider, pollution is sizable, which should be taken into account. We further explored the possibility to measure CP-violated Higgs-top coupling, via the forward-backward asymmetry A F B for the process e + e − → ZH. The asymmetry can reach 0.7% which is comparable to the precision of cross section measurement.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.