Explore the Dataset

Using the content information from the source, we generate synthetic videos that match the target identity. We take an additional step to identify similar faces between the source and target, ensuring that the selected source/target pair have closely matching facial features.