In this paper, we propose an efficient cascaded model for sign language recognition taking benefit from spatio-temporal hand-based information using deep learning approaches, especially Single Shot Detector (SSD), Convolutional Neural Network (CNN), and Long Short Term Memory (LSTM), from videos. Our simple yet efficient and accurate model includes two main parts: hand detection and sign recognition. Three types of spatial features, including hand features, Extra Spatial Hand Relation (ESHR) features, and Hand Pose (HP) features, have been fused in the model to feed to LSTM for temporal features extraction. We train SSD model for hand detection using some videos collected from five online sign dictionaries. Our model is evaluated on our proposed dataset (Rastgoo et al., Expert Syst Appl 150: 113336, 2020), including 10’000 sign videos for 100 Persian sign using 10 contributors in 10 different backgrounds, and isoGD dataset. Using the 5-fold cross-validation method, our model outperforms state-of-the-art alternatives in sign language recognition
As an alternative to the energy-intensive evaporation-crystallization method, membrane distillation crystallization (MDC) was applied for the first time to obtain calcium nitrate crystals from its aqueous solution. Calcium nitrate solution was obtained through the reaction between calcium carbonate and nitric acid, and then it was concentrated in the membrane distillation (MD) process and further crystallized. The MD step was conducted using hydrophobic polyvinylidene fluoride (PVDF)/sorbitan trioleate (Span 85) membranes. Span 85 was incorporated into the membrane structure in various concentrations to improve the hydrophobicity of membranes, and the resultant membranes were characterized via different methods. In addition, the resultant calcium nitrate crystals were characterized by X-ray fluorescence (XRF) spectroscopy. The MDC results showed that the optimum amount of Span 85 in the polymeric solution was 4%, which led to the formation of a membrane with higher porosity (67.2%) and water contact angle (95.7°) compared to the neat PVDF membrane. The mentioned membrane exhibited the highest water flux in the MD process compared to the other membranes, and also it produced the highest amount of crystals due to its remarkably better performance in the MD step in terms of feed concentration. 相似文献
The Journal of Supercomputing - Graphics processing units (GPUs) are powerful in performing data-parallel applications. Such applications most often rely on the GPU’s memory hierarchy to... 相似文献
For the first time in this study, Zinc oxide nanoparticles were biosynthesized by the eco-friendly and cost-effective procedure using Amygdalus scoparia stem bark extract then used as antibacterial, antifungal, anticancer, and anti-diabetic agents. The characterization techniques confirmed the biosynthesis, crystalline nature, structure, size, elemental composition of ZnO NPs and bioactive compounds that exist in A. scoparia extract accounting for Zn2+ ion reduction, capping and stabilization of ZnO NPs. The ZnO NPs displayed remarkable inhibitory activity against E. coli, E. aerigenes, S. aureus, P. oryzae, F. thapsinum, and F. semitectum compared to antibiotic standards. The ZnO NPs showed significant inhibitory effects on cancer cell lines, while it had no toxic effect on Vero normal cell line. The ZnO NPs (30 mg/kg)-treated diabetic rats showed significantly higher levels of insulin and lower AST, ALT and blood glucose compared with the STZ induced diabetic group and other treated groups (P < 0.05). The ZnO NPs- and extract-treated rats showed significantly higher levels of IR, GluT2, and GCK expression and lower TNFα expression compared with the STZ induced diabetic rats. Our findings showed that ZnO NPs represented an outstanding performance for biological applications. 相似文献
Wireless Networks - Data dissemination toward static sinks causes the nearby nodes to deplete their energy quicker than the other nodes in the field (i.e., this is referred to as the hotspot... 相似文献
Novel membranes consisting of polyvinyl alcohol (PVA), cellulose nanocrystals (CNCs) originated from Rice Husk (RH), Glutaraldehyde (GLA) and Glycerine (G) were manufactured by the compression moulding process. Rice husk is a new source to isolate pure cellulose nanocrystals via mechanical and chemical treatment. The biodegradability of the membranes has been evaluated using UV accelerated weathering as well as a soil burial test. The morphology of membranes (PVA/RH-CNCs) was characterized by using the Scanning Electron Microscope (SEM). The chemical structure of the membranes (PVA/RH-CNCs) was characterised by Fourier-Transformation Infrared (FTIR) spectroscopy as well as wide-angle X-ray diffraction. The mechanical properties of the membranes were evaluated using standard techniques. Swelling and weight loss resulting from biodegradation were also evaluated. The results showed that the developed transient membranes can be used as food packaging bags owing to biodegradability (weight loss) under irradiation and during soil burial. 相似文献
The present work investigates an active waste heat recovery system for the side walls of the aluminium electrolysis cells, enabling utilization of the extracted heat in power generation. This will potentially lead to energy efficiency improvement in the primary aluminium production industry and an enhanced aluminium production rate. An experimentally validated loop thermosyphon heat pipe model was used for heat extraction from the cell side wall. Boosting system thermal efficiency through waste heat recovery, by means of a heat utilization system, and increasing the level of control, as well as thermal equilibrium, stand as the main addressed objectives of the current study, which consequently result in an increased aluminium production rate. An organic Rankine cycle is incorporated into the system, and its performance is evaluated, taking into consideration the operating situations in terms of available temperature and thermal power range. 相似文献