Example-based texture synthesis has been an active research problem for over two decades. Still, synthesizing textures with nonlocal structures remains a challenge. In this article, we present a texture synthesis technique that builds upon convolutional neural networks and extracted statistics of pretrained deep features. We introduce a structural energy, based on correlations among deep features, which capture the self-similarities and regularities characterizing the texture. Specifically, we show that our technique can synthesize textures that have structures of various scales, local and nonlocal, and the combination of the two.

No abstract

No abstract

Severe visibility limitations resulting from fog may lead to acute transportation accidents and high losses of property and lives. Thus, reliable monitoring facilities are of extreme importance. Nevertheless, current monitoring instruments suffer from low spatial resolution, high costs, or lack of precision at near-surface levels. It has, however, recently been shown that the commercial microwave links that form the infrastructure of cellular communication networks can provide crucial information regarding the appearance of dense fog and its intensity. Typical microwave systems currently in operation make use of frequencies between 6 and 40 GHz and, thus, can only monitor heavy fog. However, there is a growing demand for high data rates and expanded bandwidth in modern mobile radio networks. As a result, higher frequencies (e.g., around 80 GHz) are being implemented in order to fulfill these increased requirements. Notably, the attenuation induced as a result of fog at a given intensity increases as operating frequency rises, allowing, for the first time, the possibility of using this system to monitor typical fog intensities, at high resolution and low cost. Here, a theoretical simulation is presented in which simulated fog patches are introduced into an area where a network of links is deployed. Two-dimensional maps are generated utilizing the simulated microwave network to represent sensitivity thresholds for fog detection at three different frequencies: 20, 38, and 80 GHz. Real-data measurements of fog are also demonstrated using 38-GHz band links. The results indicate the vast future potential of commercial microwave links as an opportunistic system for monitoring fog.

No abstract