RESUMO
Pedestrian detection based on deep learning methods have reached great success in the past few years with several possible real-world applications including autonomous driving, robotic navigation, and video surveillance. In this work, a new neural network two-stage pedestrian detector with a new custom classification head, adding the triplet loss function to the standard bounding box regression and classification losses, is presented. This aims to improve the domain generalization capabilities of existing pedestrian detectors, by explicitly maximizing inter-class distance and minimizing intra-class distance. Triplet loss is applied to the features generated by the region proposal network, aimed at clustering together pedestrian samples in the features space. We used Faster R-CNN and Cascade R-CNN with the HRNet backbone pre-trained on ImageNet, changing the standard classification head for Faster R-CNN, and changing one of the three heads for Cascade R-CNN. The best results were obtained using a progressive training pipeline, starting from a dataset that is further away from the target domain, and progressively fine-tuning on datasets closer to the target domain. We obtained state-of-the-art results, MR-2 of 9.9, 11.0, and 36.2 for the reasonable, small, and heavy subsets on the CityPersons benchmark with outstanding performance on the heavy subset, the most difficult one.
RESUMO
Acute and chronic dermatological injuries need rapid tissue repair due to the susceptibility to infections. To effectively promote cutaneous wound recovery, it is essential to develop safe, low-cost, and affordable regenerative tools. Therefore, we aimed to identify the biological mechanisms involved in the wound healing properties of the glycosaminoglycan dermatan sulfate (DS), obtained from ascidian Styela plicata, a marine invertebrate, which in preliminary work from our group showed no toxicity and promoted a remarkable fibroblast proliferation and migration. In this study, 2,4-DS (50 µg/mL)-treated and control groups had the relative gene expression of 84 genes participating in the healing pathway evaluated. The results showed that 57% of the genes were overexpressed during treatment, 16% were underexpressed, and 9.52% were not detected. In silico analysis of metabolic interactions exhibited overexpression of genes related to: extracellular matrix organization, hemostasis, secretion of inflammatory mediators, and regulation of insulin-like growth factor transport and uptake. Furthermore, in C57BL/6 mice subjected to experimental wounds treated with 0.25% 2,4-DS, the histological parameters demonstrated a great capacity for vascular recovery. Additionally, this study confirmed that DS is a potent inducer of wound-healing cellular pathways and a promoter of neovascularization, being a natural ally in the tissue regeneration strategy.
Assuntos
Dermatan Sulfato , Urocordados , Animais , Camundongos , Dermatan Sulfato/metabolismo , Dermatan Sulfato/farmacologia , Camundongos Endogâmicos C57BL , Urocordados/metabolismo , Cicatrização , Recursos NaturaisRESUMO
La medicina en general y muy en especial la cardiología, han sentido el impacto directo de las nuevas posibilidades abiertas por el concepto "integración en muy alta escala" (VSLI), que integra microcomputadores, memorias RAM de gran tamaño y otros. Los autores analizan aplicaciones importantes del desarrollo ingenieril electrónico en el campo de la cardiología, como son los marcapasos inteligentes, los desfibriladores automáticos implantables, el monitoreo electrocardiográfico Holter y procesamiento de información clínica obtenida en el test de esfuerzo