Arxiv April 20, 2018 Existing deep learning-based image inpainting methods often lead to artifacts such as color discrepancy and blurriness. Researchers in the US (NVIDIA Corporation) propose the use of partial convolutions, where the convolution is masked and renormalized to be conditioned on only valid pixels. They include a mechanism to automatically generate an updated mask for the next layer as part of the forward pass. They have demonstrated that their model outperforms other methods for irregular masks. They have shown qualitative and quantitative comparisons with other methods to validate their approach… read more. Open Access TECHNICAL ARTICLE
Tag Archives: Imaging technology
Face recognition technology that works in the dark
Science Daily April 16, 2018 Researchers at the U.S. Army Research Laboratory used advanced domain adaptation techniques based on deep neural networks to overcome the challenge of cross-spectrum, or heterogeneous, face recognition to process thermal image of a person’s face captured in low-light or nighttime conditions. The fundamental approach is composed of two key parts: a non-linear regression model that maps a given thermal image into a corresponding visible latent representation and an optimization problem that projects the latent projection back into the image space. In face verification experiments using a common open source deep neural network architecture, their approach […]
New algorithm can create movies from just a few snippets of text
Science Magazine February 23, 2018 Researchers in Belgium have developed a machine learning a neural network algorithm. During training, software assesses its performance after each attempt, and feedback circulates through the millions of network connections to refine future computations. The first stage uses the text to create a “gist” of the video, the second stage takes both the gist and the text and produces a short video. During training, a second network acts as a “discriminator.” As it gets better, it becomes a harsher critic, and its feedback sets a higher bar for the generator network. Currently, the videos are […]
Using technology to detect hidden threats
Eurekalert February 9, 2018 A team of researchers in the US (University of Delaware, Army Research Laboratory) is creating an augmented reality system that will use traditional cameras, thermal infrared sensing, and ground penetrating radar to find and classify potentially dangerous objects from up to 30 meters away. The radar can detect objects buried up to 3-5 inches. The multi-camera system can be used in the night, foggy conditions, and dust storms. It can be deployed on autonomous vehicles, drones, or robots sent to scout the surroundings before troops move in. The technology is being tested in a training facility… […]