In this paper, we tackle Automatic Meter Reading (AMR) by leveraging the high capability of Convolutional Neural Networks (CNNs). We design a two-stage approach that employs the Fast-YOLO object detector for counter detection and evaluates three different CNN-based approaches for counter recognition. In the AMR literature, most datasets are not available to the research community since the images belong to a service company. In this sense, we introduce a new public dataset, called UFPR-AMR dataset, with 2,000 fully and manually annotated images. This dataset is, to the best of our knowledge, three times larger than the largest public dataset found in the literature and contains a well-defined evaluation protocol to assist the development and evaluation of AMR methods. Furthermore, we propose the use of a data augmentation technique to generate a balanced training set with many more examples to train the CNN models for counter recognition. In the proposed dataset, impressive results were obtained and a detailed speed/accuracy trade-off evaluation of each model was performed. In a public dataset, state-of-the-art results were achieved using less than 200 images for training.
Studying the strength of ties in social networks allows to identify impact at micro-macro levels in the network, to analyze how distinct relationships play different roles, and so on. Indeed, the strength of ties has been investigated in many contexts with different goals. Here, we aim to address the problem of measuring ties strength in co-authorship social networks. Specifically, we present four case studies detailing problems with current metrics and propose a new one. Then, we build a co-authorship social network by using a real digital library and identify how the strength of ties relates to the quality of publication venues when measured by different topological properties. Our results show the best ranked venues have similar patterns of strength of co-authorship ties.
Studies have analyzed social networks considering a plethora of metrics for different goals, from improving e-learning to recommend people and things. Here, we focus on large-scale social networks defined by researchers and their common published articles, which form co-authorship social networks. Then, we introduce CNARe, an online tool that analyzes the networks and present recommendations of collaborations based on three different algorithms (Affin, CORALS and MVCWalker). Through visualizations and social networks metrics, CNARe also allows to investigate how the recommendations affect the co-authorship social networks, how researchers' networks are in a central and eagle-eye context, and how the strength of ties behaves in large co-authorship social networks. Furthermore, users can upload their own network in CNARe and make their own recommendation and social network analysis.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.