Boundless: Generative Adversarial Networks for Image Extension
This addresses a specific challenge in image editing and computational photography, offering an incremental improvement over prior inpainting techniques.
The paper tackled the problem of image extension, where existing inpainting methods produce blurry or repetitive results, by introducing semantic conditioning to a GAN discriminator, achieving coherent semantics and visually pleasing outputs, including in extreme cases like panorama generation.
Image extension models have broad applications in image editing, computational photography and computer graphics. While image inpainting has been extensively studied in the literature, it is challenging to directly apply the state-of-the-art inpainting methods to image extension as they tend to generate blurry or repetitive pixels with inconsistent semantics. We introduce semantic conditioning to the discriminator of a generative adversarial network (GAN), and achieve strong results on image extension with coherent semantics and visually pleasing colors and textures. We also show promising results in extreme extensions, such as panorama generation.