Image Outpainting: Hallucinating Beyond the Image

Abstract

Image inpainting is a technique that aims to fill in the missing regions with visually plausible content. However, an opposite idea, which is painting outside images, receives little work. In this study, we investigate the issue of image outpainting. Considering that the model needs better prediction ability as there is less neighboring information in image outpainting, the study proposes a novel image outpainting architecture that can obtain both deep model performance and detailed information. To fully take advantage of residual learning, dense residual (DR) learning is proposed and the image generative network is built on DR. To avoid losing subtle information caused by downsampling in encoder-decoder, shortcuts are added for transferring previous knowledge. Different from vanilla U-Net, we propose a skip method of the semicomplete form. Experimental results show that the proposed method achieves excellent performance.

Doom or Deliciousness: Challenges and Opportunities for Visualization in the Age of Generative Models

“…In addition to prompts, participants were invited to create their own and thereby ruminate on their concerns and aspirations for these tools. These discussions were interleaved with situating prompts that reviewed various functionalities possessed by current tools (such as in‐painting [ YQS20 ] and out‐painting [ XLC20 ]) to help situate their thinking among current technologies. We utilized this shared creative space to invite thinking aloud and ideation, which may have been more limited in a more restricted environment.…”

Street-View Image Generation from a Bird's-Eye View Layout

“…However, this approach can be computationally expensive and thus is limited to generating low-resolution images. Subsequently, GANs were introduced to the task [16,23,39,49] which do not suffer from the same computational limitations as pixel-wise autoregressive approaches. More recent works have utilized a Vector Quantised-Variational Autoencoder (VQ-VAE) [47] to great success [2,4].…”

The building plan automatic design optimization based on CFD numerical intelligent simulation

“…In the setting of optimization, according to documents, 37 the smaller population would have a better performance; therefore, 300 times of assessment was abundant enough to get excellent results. As the population size became smaller, to avoid running into the local optimization solution, the mating rate and mutation rate would need to be enhanced.…”