Abstract

DeepFaceLab is an open-source deepfake system created by iperov for face swapping with more than 3,000 forks and 14,000 stars in Github: it provides an imperative and easy-to-use pipeline for people to use with no comprehensive understanding of deep learning framework or with model implementation required, while remains a flexible and loose coupling structure for people who need to strengthen their own pipeline with other features without writing complicated boilerplate code. In this paper, we detail the principles that drive the implementation of DeepFaceLab and introduce the pipeline of it, through which every aspect of the pipeline can be modified painlessly by users to achieve their customization purpose, and it's noteworthy that DeepFaceLab could achieve results with high fidelity and indeed indiscernible by mainstream forgery detection approaches. We demonstrate the advantage of our system through comparing our approach with current prevailing systems. 1

No abstract

Background Only 30% or fewer of individuals at clinical high risk (CHR) convert to full psychosis within 2 years. Efforts are thus underway to refine risk identification strategies to increase their predictive power. Our objective was to develop and validate the predictive accuracy and individualized risk components of a mobile app-based psychosis risk calculator (RC) in a CHR sample from the SHARP (ShangHai At Risk for Psychosis) program. Method In total, 400 CHR individuals were identified by the Chinese version of the Structured Interview for Prodromal Syndromes. In the first phase of 300 CHR individuals, 196 subjects (65.3%) who completed neurocognitive assessments and had at least a 2-year follow-up assessment were included in the construction of an RC for psychosis. In the second phase of the SHARP sample of 100 subjects, 93 with data integrity were included to validate the performance of the SHARP-RC. Results The SHARP-RC showed good discrimination of subsequent transition to psychosis with an AUC of 0.78 (p < 0.001). The individualized risk generated by the SHARP-RC provided a solid estimation of conversion in the independent validation sample, with an AUC of 0.80 (p = 0.003). A risk estimate of 20% or higher had excellent sensitivity (84%) and moderate specificity (63%) for the prediction of psychosis. The relative contribution of individual risk components can be simultaneously generated. The mobile app-based SHARP-RC was developed as a convenient tool for individualized psychosis risk appraisal. Conclusions The SHARP-RC provides a practical tool not only for assessing the probability that an individual at CHR will develop full psychosis, but also personal risk components that might be targeted in early intervention.

No abstract