Image Outpainting and Harmonization using Generative Adversarial Networks

Although the inherently ambiguous task of predicting what resides beyond all edges of an image has rarely been explored before, we demonstrate that hold powerful potential in producing reasonable extrapolations. Two methods are proposed that aim to instigate this line of research: first approach uses a context encoder inspired by common inpainting and paradigms, while the second approach adds an extra-processing step using a single-image generative model. This way, the details are integrated with the style of the original image, in an to further boost the quality of the result and possibly allow for output resolutions to be supported.