Face beautification: Beyond makeup transfer

Facial appearance plays an important role in our social lives. Subjective perception of women's beauty depends on various face-related (e.g., skin, shape, hair) and environmental makeup, lighting, angle) factors. Similarly to cosmetic surgery the physical world, virtual face beautification is emerging field with many open issues be addressed. Inspired by latest advances style-based synthesis prediction, we propose a novel framework for beautification. For given reference high score, GAN-based architecture capable translating inquiry into sequence beautified images referenced style target score values. To achieve this objective, integrate both representation (extracted from face) prediction (trained SCUT-FBP database) process. Unlike makeup transfer, approach targets many-to-many (instead one-to-one) translation, where multiple outputs can defined different references scores. Extensive experimental results are reported demonstrate effectiveness flexibility proposed framework. support reproducible research, source codes accompanying work will made publicly available GitHub.