Universidade Federal de Minas Gerais

About: Universidade Federal de Minas Gerais is a education organization based out in Belo Horizonte, Minas Gerais, Brazil. It is known for research contribution in the topics: Population & Context (language use). The organization has 41631 authors who have published 75688 publications receiving 1249905 citations.

Do deep features generalize from everyday objects to remote sensing and aerial scenes domains

Abstract: In this paper, we evaluate the generalization power of deep features (ConvNets) in two new scenarios: aerial and remote sensing image classification. We evaluate experimentally ConvNets trained for recognizing everyday objects for the classification of aerial and remote sensing images. ConvNets obtained the best results for aerial images, while for remote sensing, they performed well but were outperformed by low-level color descriptors, such as BIC. We also present a correlation analysis, showing the potential for combining/fusing different ConvNets with other descriptors or even for combining multiple ConvNets. A preliminary set of experiments fusing ConvNets obtains state-of-the-art results for the well-known UCMerced dataset.

The ancestry of Brazilian mtDNA lineages.

Abstract: We have analyzed 247 Brazilian mtDNAs for hypervariable segment (HVS)–I and selected restriction fragment-length–polymorphism sites, to assess their ancestry in different continents. The total sample showed nearly equal amounts of Native American, African, and European matrilineal genetic contribution but with regional differences within Brazil. The mtDNA pool of present-day Brazilians clearly reflects the imprints of the early Portuguese colonization process (involving directional mating), as well as the recent immigrant waves (from Europe) of the last century. The subset of 99 mtDNAs from the southeastern region encompasses nearly all mtDNA haplogroups observed in the total Brazilian sample; for this regional subset, HVS-II was analyzed, providing, in particular, some novel details of the African mtDNA phylogeny.

Spermatogenesis and cycle of the seminiferous epithelium.

Abstract: Spermatogenesis is a complex biological process of cellular transformation that produces male haploid germ cells from diploid spermatogonial stem cells. This process has been simplified morphologically by recognizing cellular associations or ‘stages’ and ‘phases’ of spermatogenesis, which progress through precisely timed and highly organized cycles. These cycles of spermatogenesis are essential for continuous sperm production, which is dependent upon numerous factors, both intrinsic (Sertoli and germ cells) and extrinsic (androgens, retinoic acids), as well as being species-specific.

Localization systems for wireless sensor networks

Abstract: Monitoring applications define an important class of applications used in wireless sensor networks In these applications the network perceives the environment and searches for event occurrences (phenomena) by sensing different physical properties, such as temperature, humidity, pressure, ambient light, movement, and presence (for target tracking) In such cases the location information of both phenomena and nodes is usually required for tracking and correlation purposes In this work we summarize most of the concepts related to localization systems for WSNs as well as how to localize the nodes in these networks (which allows the localization of phenomena) By dividing the localization systems into three distinct components -distance/angle estimation, position computation, and localization algorithm - besides providing a didactic viewpoint, we show that these components can be seen as subareas of the localization problem that need to be analyzed and studied separately