This paper proposes the novel Pose Guided Person Generation Network (PG$^2$2)that allows to synthesize person images in arbitrary poses, based on an imageof that person and a novel pose. Our generation framework PG$^2$2 utilizes thepose information explicitly and consists of two key stages: pose integrationand image refinement. In the first stage the condition image and the targetpose are fed into a U-Net-like network to generate an initial but coarse imageof the person with the target pose. The second stage then refines the initialand blurry result by training a U-Net-like generator in an adversarial way.Extensive experimental results on both 128$\times$×64 re-identification imagesand 256$\times$×256 fashion photos show that our model generates high-qualityperson images with convincing details.