This is the implementation of FSFM-3C, a self-supervised pre-training framework to learn a transferable facial representation that boosts face security tasks.

As a classic statistical model of 3D facial shape and texture, 3D Morphable Model (3DMM) is widely used in facial analysis, e.g., model fitting, image synthesis. Conventional 3DMM is learned from a ...

An accurate computational model for image quality assessment (IQA) benefits many vision applications, such as image filtering, image processing, and image generation. Although the study of face images ...

As a beauty editor who barely ever moisturises below my neck, I asked skin experts how often I should be moisturising my face and body - and this is what they said ...