![](/assets/icon.png)
1. This paper proposes a Pose and Expression Robust Spatial-Aware GAN (PSGAN) for makeup transfer between images with large pose and expression differences.
2. PSGAN utilizes Makeup Distill Network to disentangle the makeup of the reference image as two spatial-aware makeup matrices, and an Attentive Makeup Morphing module to specify how the makeup of a pixel in the source image is morphed from the reference image.
3. A dataset containing facial images with various poses and expressions was collected for evaluations.
The article is generally trustworthy and reliable, as it provides detailed information about the proposed method, PSGAN, which has been tested on a dataset of facial images with various poses and expressions. The authors have also provided evidence for their claims by citing relevant research papers in the field. However, there are some potential biases that should be noted. For example, the authors do not discuss any possible risks associated with using this technology or any potential ethical implications of its use. Additionally, they do not explore any counterarguments or present both sides equally when discussing their proposed method. Finally, there may be some promotional content in the article as it is written by researchers who developed PSGAN themselves.