Abstract: Pose-guided person image generation and animation aim to transform a source person image to target poses. These tasks require spatial manipulation of source data. However, Convolutional ...