Ryerson University

About: Ryerson University is a education organization based out in Toronto, Ontario, Canada. It is known for research contribution in the topics: Computer science & Population. The organization has 7671 authors who have published 20164 publications receiving 394976 citations. The organization is also known as: Ryerson Polytechnical Institute & Ryerson Institute of Technology.

Back to Basics: Unsupervised Learning of Optical Flow via Brightness Constancy and Motion Smoothness

Abstract: Recently, convolutional networks (convnets) have proven useful for predicting optical flow. Much of this success is predicated on the availability of large datasets that require expensive and involved data acquisition and laborious labeling. To bypass these challenges, we propose an unsupervised approach (i.e., without leveraging groundtruth flow) to train a convnet end-to-end for predicting optical flow between two images. We use a loss function that combines a data term that measures photometric constancy over time with a spatial term that models the expected variation of flow across the image. Together these losses form a proxy measure for losses based on the groundtruth flow. Empirically, we show that a strong convnet baseline trained with the proposed unsupervised approach outperforms the same network trained with supervision on the KITTI dataset.

Sparseness Meets Deepness: 3D Human Pose Estimation from Monocular Video

Abstract: This paper addresses the challenge of 3D full-body human pose estimation from a monocular image sequence. Here, two cases are considered: (i) the image locations of the human joints are provided and (ii) the image locations of joints are unknown. In the former case, a novel approach is introduced that integrates a sparsity-driven 3D geometric prior and temporal smoothness. In the latter case, the former case is extended by treating the image locations of the joints as latent variables to take into account considerable uncertainties in 2D joint locations. A deep fully convolutional network is trained to predict the uncertainty maps of the 2D joint locations. The 3D pose estimates are realized via an Expectation-Maximization algorithm over the entire sequence, where it is shown that the 2D joint location uncertainties can be conveniently marginalized out during inference. Empirical evaluation on the Human3.6M dataset shows that the proposed approaches achieve greater 3D pose estimation accuracy over state-of-the-art baselines. Further, the proposed approach outperforms a publicly available 2D pose estimation baseline on the challenging PennAction dataset.

Urban ecosystems and the North American carbon cycle

Abstract: Approximately 75–80% of the population of North America currently lives in urban areas as defined by national census bureaus, and urbanization is continuing to increase. Future trajectories of fossil fuel emissions are associated with a high degree of uncertainty; however, if the activities of urban residents and the rate of urban land conversion can be captured in urban systems models, plausible emissions scenarios from major cities may be generated. Integrated land use and transportation models that simulate energy use and traffic-related emissions are already in place in many North American cities. To these can be added a growing dataset of carbon gains and losses in vegetation and soils following urbanization, and a number of methods of validating urban carbon balance modeling, including top down atmospheric monitoring and urban ‘metabolic’ studies of whole ecosystem mass and energy flow. Here, we review the state of our understanding of urban areas as whole ecosystems with regard to carbon balance, including both drivers of fossil fuel emissions and carbon cycling in urban plants and soils. Interdisciplinary, whole-ecosystem studies of the socioeconomic and biophysical factors that influence urban carbon cycles in a range of cities may greatly contribute to improving scenarios of future carbon balance at both continental and global scales.

Social Alliances: Company/Nonprofit Collaboration

Abstract: Companies are increasingly seeing corporate social responsibility as a key to long-term success and are collaborating with nonprofit organizations in various ways to establish themselves as good corporate citizens. This article delves into a promising form of company/nonprofit collaboration called social alliances, which are long-term, collaborative efforts between companies and nonprofits that are designed to achieve strategic objectives for both organizations. The characteristics, factors, and circumstances that enable or impede social alliances are examined through an investigation of 11 social alliances involving 26 organizations. Though social alliances may be fraught with problems, they can be designed, structured, nurtured, and maintained in a manner that will enable them both to contribute to solving pressing social problems and to fulfilling important strategic objectives for companies and nonprofits.