Explore the vast range of titles available.

Recent years have witnessed the popularity of using two-stream convolutional neural networks for action recognition. However, existing two-stream convolutional neural network-based action recognition approaches are incapable of distinguishing some roughly similar actions in videos such as sneezing and yawning. To solve this problem, we propose a Multi-head Attention-based Two-stream EfficientNet (MAT-EffNet) for action recognition, which can take advantage of the efficient feature extraction of EfficientNet. The proposed network consists of two streams (i.e., a spatial stream and a temporal stream), which first extract the spatial and temporal features from consecutive frames by using EfficientNet. Then, a multi-head attention mechanism is utilized on the two streams to capture the key action information from the extracted features. The final prediction is obtained via a late average fusion, which averages the softmax score of spatial and temporal streams. The proposed MAT-EffNet can focus on the key action information at different frames and compute the attention multiple times, in parallel, to distinguish similar actions. We test the proposed network on the UCF101, HMDB51 and Kinetics-400 datasets. Experimental results show that the MAT-EffNet outperforms other state-of-the-art approaches for action recognition.

Understanding the spatio‐temporal dynamics of runoff generation in headwater catchments is challenging, due to the intermittent and fragmented nature of surface flows. The active stream network in non‐perennial rivers contracts and expands, with a dynamic behavior that depends on the complex interplay among climate, topography, and geology. In this work, CATchment HYdrology, an integrated surface–subsurface hydrological model (ISSHM), is used to simulate the stream network dynamics of two virtual catchments with the same, spatially homogeneous, subsurface characteristics (hydraulic conductivity, porosity, water retention curves) but different morphology. We run two sets of simulations to reproduce a sequence of steady‐states at different catchment wetness levels and transient conditions and analyze the joint variations of the stream length (L) and discharge at the outlet (Q) with high spatio‐temporal resolutions. The shape of the L(Q) curves differs in the two catchments but does not depend on the climate forcing, as it is mainly controlled by the underlying topography. We then analyzed the suitability of the topographic wetness index and the contributing area to identify the spatial configuration of the maximum stream length in the two catchments. These two morphometric parameters provided a good estimate of the spatial distribution of the maximum flowing network in both the study catchments. Our numerical simulations indicate that ISSHMs have the potential to accurately describe the spatio‐temporal variations of the stream networks and the processes driving such dynamic behavior and that, overall, they can be useful tools to gain insights into the main physical drivers of non‐perennial streams. Key Points A physics‐based model is used to reproduce the spatiotemporal stream network dynamics of two virtual catchments with different morphology The link between active stream length and outlet discharge is analyzed in steady‐state and transient conditions Well‐established topographic indices can identify the maximum active network extent in catchments with homogeneous subsurface properties

The integration of solar energy with a power system brings great economic and environmental benefits. However, the high penetration of solar power is challenging due to the operation and planning of the existing power system owing to the intermittence and randomicity of solar power generation. Achieving accurate predictions for power generation is important to provide high-quality electric energy for end-users. Therefore, in this paper, we introduce a deep learning-based dual-stream convolutional neural network (CNN) and long short-term nemory (LSTM) network followed by a self-attention mechanism network (DSCLANet). Here, CNN is used to learn spatial patterns and LSTM is incorporated for temporal feature extraction. The output spatial and temporal feature vectors are then fused, followed by a self-attention mechanism to select optimal features for further processing. Finally, fully connected layers are incorporated for short-term solar power prediction. The performance of DSCLANet is evaluated on DKASC Alice Spring solar datasets, and it reduces the error rate up to 0.0136 MSE, 0.0304 MAE, and 0.0458 RMSE compared to recent state-of-the-art methods.

While temporary streams account for more than half of the global discharge, high spatiotemporal resolution data on the three main hydrological states (dry streambed, standing water, and flowing water) of temporary stream remains sparse. This study presents a low-cost, multi-sensor system to monitor the hydrological state of temporary streams in mountainous headwaters. The monitoring system consists of an Arduino microcontroller board combined with an SD-card data logger shield, and four sensors: an electrical resistance (ER) sensor, temperature sensor, float switch sensor, and flow sensor. The monitoring system was tested in a small mountainous headwater catchment, where it was installed on multiple locations in the stream network, during two field seasons (2016 and 2017). Time-lapse cameras were installed at all monitoring system locations to evaluate the sensor performance. The field tests showed that the monitoring system was power efficient (running for nine months on four AA batteries at a five-minute logging interval) and able to reliably log data (<1% failed data logs). Of the sensors, the ER sensor (99.9% correct state data and 90.9% correctly timed state changes) and flow sensor (99.9% correct state data and 90.5% correctly timed state changes) performed best (2017 performance results). A setup of the monitoring system with these sensors can provide long-term, high spatiotemporal resolution data on the hydrological state of temporary streams, which will help to improve our understanding of the hydrological functioning of these important systems.

Modelling the effects of environmental change on populations is a key challenge for ecologists, particularly as the pace of change increases. Currently, modelling efforts are limited by difficulties in establishing robust relationships between environmental drivers and population responses. We developed an integrated capture–recapture state‐space model to estimate the effects of two key environmental drivers (stream flow and temperature) on demographic rates (body growth, movement and survival) using a long‐term (11 years), high‐resolution (individually tagged, sampled seasonally) data set of brook trout (Salvelinus fontinalis) from four sites in a stream network. Our integrated model provides an effective context within which to estimate environmental driver effects because it takes full advantage of data by estimating (latent) state values for missing observations, because it propagates uncertainty among model components and because it accounts for the major demographic rates and interactions that contribute to annual survival. We found that stream flow and temperature had strong effects on brook trout demography. Some effects, such as reduction in survival associated with low stream flow and high temperature during the summer season, were consistent across sites and age classes, suggesting that they may serve as robust indicators of vulnerability to environmental change. Other survival effects varied across ages, sites and seasons, indicating that flow and temperature may not be the primary drivers of survival in those cases. Flow and temperature also affected body growth rates; these responses were consistent across sites but differed dramatically between age classes and seasons. Finally, we found that tributary and mainstem sites responded differently to variation in flow and temperature. Annual survival (combination of survival and body growth across seasons) was insensitive to body growth and was most sensitive to flow (positive) and temperature (negative) in the summer and fall. These observations, combined with our ability to estimate the occurrence, magnitude and direction of fish movement between these habitat types, indicated that heterogeneity in response may provide a mechanism providing potential resilience to environmental change. Given that the challenges we faced in our study are likely to be common to many intensive data sets, the integrated modelling approach could be generally applicable and useful.

The utilization of surface water isoscapes facilitates the characterization of source‐water contributions and hydrological processes within a basin. However, intricate river network topology poses significant challenges in applying this approach to large and complex basins. In this study, a spatial stream network model (SSNM) was employed to create isoscapes of surface waters using 852 river isotopic data across the Yangtze River Basin (YRB). Results showed that precipitation and river water δ18O values displayed a similar trend, characterized by the lowest (highest) values in the upper (lower) reaches of the YRB. River water δ18O exhibited multiple spatial dependencies regarding the flow‐connected, flow‐unconnected, and Euclidean spatial relationships from the shape of semivariograms, indicating basin processes within the river network and terrestrial landscape across the YRB. The riverine δ18O isoscapes were predicted by coupling environmental covariates including hydrologic, climatic drivers, and landscapes with spatial autocovariance structures across the YRB. The predictive accuracy of isoscapes from SSNM was distinctly improved from 66% using linear model to 87%. Maps of source contributions from SSNM showed higher contribution from meltwater of glacier/permafrost in the upper reaches of YRB (>60%). In contrast, precipitation and groundwater were the main contributing recharge sources due to the distributed aquifer and evaporative effects on river water in the mid‐lower reaches of YRB. The findings present a novel approach for the representation of isoscapes in large‐scale, intricate basins, offering valuable evidence for provenance studies and basin management.

Species distribution models (SDMs) are frequently data‐limited. In aquatic habitats, emerging environmental DNA (eDNA) sampling methods can be quicker and more cost‐efficient than traditional count and capture surveys, but their utility for fitting SDMs is complicated by dilution, transport, and loss processes that modulate DNA concentrations and mix eDNA from different locations. Past models for estimating organism densities from measured species‐specific eDNA concentrations have accounted for how these processes affect expected concentrations. We built off this previous work to construct a linear hierarchical model that also accounts for how they give rise to spatially correlated concentration errors. We applied our model to 60 simulated stream networks and three types of species niches in order to answer two questions: 1) what is the D‐optimal sampling design, i.e. where should eDNA samples be positioned to most precisely estimate species–environment relationships? and 2) How does parameter estimation accuracy depend on the stream network's topological and hydrologic properties? We found that correcting for eDNA dynamics was necessary to obtain consistent parameter estimates, and that relative to a heuristic benchmark design, optimizing sampling locations improved design efficiency by an average of 41.5%. Samples in the D‐optimal design tended to be positioned near downstream ends of stream reaches high in the watershed, where eDNA concentration was high and mostly from homogeneous source areas, and they collectively spanned the full ranges of covariates. When measurement error was large, it was often optimal to collect replicate samples from high‐information reaches. eDNA‐based estimates of species–environment regression parameters were most precise in stream networks that had many reaches, large geographic size, slow flows, and/or high eDNA loss rates. Our study demonstrates the importance and viability of accounting for eDNA dilution, transport, and loss in order to optimize sampling designs and improve the accuracy of eDNA‐based species distribution models.

Human activity recognition (HAR) is pivotal in advancing applications ranging from healthcare monitoring to interactive gaming. Traditional HAR systems, primarily relying on single data sources, face limitations in capturing the full spectrum of human activities. This study introduces a comprehensive approach to HAR by integrating two critical modalities: RGB imaging and advanced pose estimation features. Our methodology leverages the strengths of each modality to overcome the drawbacks of unimodal systems, providing a richer and more accurate representation of activities. We propose a two-stream network that processes skeletal and RGB data in parallel, enhanced by pose estimation techniques for refined feature extraction. The integration of these modalities is facilitated through advanced fusion algorithms, significantly improving recognition accuracy. Extensive experiments conducted on the UTD multimodal human action dataset (UTD MHAD) demonstrate that the proposed approach exceeds the performance of existing state-of-the-art algorithms, yielding improved outcomes. This study not only sets a new benchmark for HAR systems but also highlights the importance of feature engineering in capturing the complexity of human movements and the integration of optimal features. Our findings pave the way for more sophisticated, reliable, and applicable HAR systems in real-world scenarios.

Action recognition techniques based on skeleton data are receiving more and more attention in the field of computer vision due to their ability to adapt to dynamic environments and complex backgrounds. Topologizing human skeleton data as spatial-temporal graphs and processing them using graph convolutional networks (GCNs) has been shown to produce good recognition results. However, with existing GCN methods, a fixed-size convolution kernel is often used to extract time-domain features, which may not be very suitable for multi-level model structures. Equal proportion fusion of different streams in a multi-stream network may ignore the difference in recognition ability of different streams, and these will affect the final recognition result. In this paper, we are proposing (1) a multi-scale dilated temporal graph convolution layer (MDTGCL) and (2) a multi-branch feature fusion (MFF) structure. The MDTGCL utilizes multiple convolution kernels and dilated convolution to better adapt to the multi-layer structure of the GCN model and to obtain longer periods of contextual spatial-temporal information, resulting in richer behavioural features. MFF entails weighted fusion based on the results of multi-stream outputs, and this is used to obtain the final recognition results. As higher-order skeleton data are highly discriminative and more conducive to human action recognition, we used spatial information on joints and bones and their multiple motion, as well as angle information pertaining to bones, to model together in this study. By combining the above, we designed a multi-stream, multi-scale dilated spatial-temporal graph convolutional network (2M-STGCN) model and conducted extensive experiments with two large datasets (NTU RGB+D 60 and Kinetics Skeleton 400), which showed that our model performs at SOTA level.