Adaptive temporal compression for reduction of computational complexity in human behavior recognition.

The research on video analytics especially in the area of human behavior recognition has become increasingly popular recently. It is widely applied in virtual reality, video surveillance, and video retrieval. With the advancement of deep learning algorithms and computer hardware, the conventional two-dimensional convolution technique for training video models has been replaced by three-dimensional convolution, which enables the extraction of spatio-temporal features. Specifically, the use of 3D convolution in human behavior recognition has been the subject of growing interest. However, the increased dimensionality has led to challenges such as the dramatic increase in the number of parameters, increased time complexity, and a strong dependence on GPUs for effective spatio-temporal feature extraction. The training speed can be considerably slow without the support of powerful GPU hardware. To address these issues, this study proposes an Adaptive Time Compression (ATC) module. Functioning as an independent component, ATC can be seamlessly integrated into existing architectures and achieves data compression by eliminating redundant frames within video data. The ATC module effectively reduces GPU computing load and time complexity with negligible loss of accuracy, thereby facilitating real-time human behavior recognition.

Quantitative analysis of the impact of climate variability and human activities on grassland productivity of the Qilian Mountain National Park, China.

To quantitatively analyze the impact of climate variability and human activities on grassland productivity of China's Qilian Mountain National Park, this study used Carnegic-Ames-Stanford Approach model (CASA) and Integrated Vegetation model improved by the Comprehensive and Sequential Classification System (CSCS) to assess the trends of grassland NPP from 2000 to 2015, the residual trend analysis method was used to quantify the impact of human activities and climate change on the grassland based on the NPP changes. The actual grassland NPP accumulation mainly occurred in June, July and August (autumn); the actual NPP showed a fluctuating upward trend with an average increase of 2.2 g C·m-2 a-1, while the potential NPP increase of 1.6 g C·m-2 a-1 and human-induced NPP decreased of 0.5 g C·m-2 a-1. The annual temperature showed a fluctuating upward trend with an average increase of 0.1°C 10a-1, but annual precipitation showed a fluctuating upward trend with an average annual increase of 1.3 mm a-1 from 2000 to 2015. The area and NPP of grassland degradation caused by climate variability was significantly greater than that caused by human activities and mainly distributed in the northwest and central regions, but area and NPP of grassland restored caused by human activities was significantly greater than that caused by climate variability and mainly distributed in the southeast regions. In conclusion, grassland in Qilian Mountain National Park showed a trend of degradation based on distribution area, but showed a trend of restoration based on actual NPP. Climate variability was the main cause of grassland degradation in the northwestern region of study area, and restoration of grassland in the eastern region was the result of the combined effects of human activities and climate variability. Under global climate change, the establishment of Qilian Mountain National Park was of great significance to the grassland's protection and the grasslands ecological restoration that have been affected by humans.

Anthropogenic Forcing of the Baltic Sea Thallium Cycle.

Anthropogenic activities have fundamentally changed the chemistry of the Baltic Sea. According to results reported in this study, not even the thallium (Tl) isotope cycle is immune to these activities. In the anoxic and sulfidic ("euxinic") East Gotland Basin today, Tl and its two stable isotopes are cycled between waters and sediments as predicted based on studies of other redox-stratified basins (e.g., the Black Sea and Cariaco Trench). The Baltic seawater Tl isotope composition (ε205Tl) is, however, higher than predicted based on the results of conservative mixing calculations. Data from a short sediment core from East Gotland Basin demonstrates that this high seawater ε205Tl value originated sometime between about 1940 and 1947 CE, around the same time other prominent anthropogenic signatures begin to appear in the same core. This juxtaposition is unlikely to be coincidental and suggests that human activities in the surrounding area have altered the seawater Tl isotope mass-balance of the Baltic Sea.

Human Activity Recognition in a Free-Living Environment Using an Ear-Worn Motion Sensor.

Human activity recognition (HAR) technology enables continuous behavior monitoring, which is particularly valuable in healthcare. This study investigates the viability of using an ear-worn motion sensor for classifying daily activities, including lying, sitting/standing, walking, ascending stairs, descending stairs, and running. Fifty healthy participants (between 20 and 47 years old) engaged in these activities while under monitoring. Various machine learning algorithms, ranging from interpretable shallow models to state-of-the-art deep learning approaches designed for HAR (i.e., DeepConvLSTM and ConvTransformer), were employed for classification. The results demonstrate the ear sensor's efficacy, with deep learning models achieving a 98% accuracy rate of classification. The obtained classification models are agnostic regarding which ear the sensor is worn and robust against moderate variations in sensor orientation (e.g., due to differences in auricle anatomy), meaning no initial calibration of the sensor orientation is required. The study underscores the ear's efficacy as a suitable site for monitoring human daily activity and suggests its potential for combining HAR with in-ear vital sign monitoring. This approach offers a practical method for comprehensive health monitoring by integrating sensors in a single anatomical location. This integration facilitates individualized health assessments, with potential applications in tele-monitoring, personalized health insights, and optimizing athletic training regimes.

A Comparative Study of Physically Accurate Synthetic Shadow Datasets in Agricultural Settings with Human Activity.

Shadow, a natural phenomenon resulting from the absence of light, plays a pivotal role in agriculture, particularly in processes such as photosynthesis in plants. Despite the availability of generic shadow datasets, many suffer from annotation errors and lack detailed representations of agricultural shadows with possible human activity inside, excluding those derived from satellite or drone views. In this paper, we present an evaluation of a synthetically generated top-down shadow segmentation dataset characterized by photorealistic rendering and accurate shadow masks. We aim to determine its efficacy compared to real-world datasets and assess how factors such as annotation quality and image domain influence neural network model training. To establish a baseline, we trained numerous baseline architectures and subsequently explored transfer learning using various freely available shadow datasets. We further evaluated the out-of-domain performance compared to the training set of other shadow datasets. Our findings suggest that AgroSegNet demonstrates competitive performance and is effective for transfer learning, particularly in domains similar to agriculture.

Anthropogenic disturbance affects calling and collective behaviour in corvid roosts.

Acoustic communication plays an important role in coordinating group dynamics and collective movements across a range of taxa. However, anthropogenic disturbance can inhibit the production or reception of acoustic signals. Here, we investigate the effects of noise and light pollution on the calling and collective behaviour of wild jackdaws (Corvus monedula), a highly social corvid species that uses vocalizations to coordinate collective movements at winter roosting sites. Using audio and video monitoring of roosts in areas with differing degrees of urbanization, we evaluate the influence of anthropogenic disturbance on vocalizations and collective movements. We found that when levels of background noise were higher, jackdaws took longer to settle following arrival at the roost in the evening and also called more during the night, suggesting that human disturbance may cause sleep disruption. High levels of overnight calling were, in turn, linked to disruption of vocal consensus decision-making and less cohesive group departures in the morning. These results raise the possibility that, by affecting cognitive and perceptual processes, human activities may interfere with animals' ability to coordinate collective behaviour. Understanding links between anthropogenic disturbance, communication, cognition and collective behaviour must be an important research priority in our increasingly urbanized world. This article is part of the theme issue 'The power of sound: unravelling how acoustic communication shapes group dynamics'.

A dataset of ambient sensors in a meeting room for activity recognition.

As IoT technology advances, using machine learning to detect user activities emerges as a promising strategy for delivering a variety of smart services. It is essential to have access to high-quality data that also respects privacy concerns and data streams from ambient sensors in the surrounding environment meet this requirement. However, despite growing interest in research, there is a noticeable lack of datasets from ambient sensors designed for public spaces, as opposed to those for private settings. To bridge this gap, we design the DOO-RE dataset within an actual meeting room environment, equipped with three types of ambient sensors: those triggered by actuators, users, and the environment itself. This dataset is compiled from the activities of over twenty students throughout a period of four months. DOO-RE provides reliable and purpose-oriented activity data in a public setting, with activity labels verified by multiple annotators through a process of cross-validation to guarantee data integrity. DOO-RE categorizes nine different types of activities and facilitates the study of both single and group activities. We are optimistic that DOO-RE will play a significant role in advancing human activity recognition technologies, enhancing smart automation systems, and enabling the rapid setup of smart spaces through ambient sensors.

Chemical Cross-Linking Cellulose Aerogel-Based Triboelectric Nanogenerators for Energy Harvesting and Sensing Human Activities.

Zinc oxide (ZnO) is a widely employed material for enhancing the performance of cellulose-based triboelectric nanogenerators (C-TENGs). Our study provides a novel chemical interpretation for the improved output efficiency of ZnO in C-TENGs. C-TENGs exhibit excellent flexibility and integration, achieving a maximum open-circuit voltage (Voc) of 210 V. The peak power density is 54.4 µW/cm2 with a load resistance of 107 Ω, enabling the direct powering of 191 light-emitting diodes with the generated electrical output. Moreover, when deployed as self-powered sensors, C-TENGs exhibit prolonged operational viability and responsiveness, adeptly discerning bending and motion induced by human interaction. The device's sensitivity, flexibility, and stability position it as a promising candidate for a diverse array of energy-harvesting applications and advanced healthcare endeavors. Specifically, envisaging sensitized wearable sensors for human activities underscores the multifaceted potential of C-TENGs in enhancing both energy-harvesting technologies and healthcare practices.

A Novel Framework Based on Deep Learning Architecture for Continuous Human Activity Recognition with Inertial Sensors.

Frameworks for human activity recognition (HAR) can be applied in the clinical environment for monitoring patients' motor and functional abilities either remotely or within a rehabilitation program. Deep Learning (DL) models can be exploited to perform HAR by means of raw data, thus avoiding time-demanding feature engineering operations. Most works targeting HAR with DL-based architectures have tested the workflow performance on data related to a separate execution of the tasks. Hence, a paucity in the literature has been found with regard to frameworks aimed at recognizing continuously executed motor actions. In this article, the authors present the design, development, and testing of a DL-based workflow targeting continuous human activity recognition (CHAR). The model was trained on the data recorded from ten healthy subjects and tested on eight different subjects. Despite the limited sample size, the authors claim the capability of the proposed framework to accurately classify motor actions within a feasible time, thus making it potentially useful in a clinical scenario.

Eco-environmental changes due to human activities in the Erhai Lake Basin from 1990 to 2020.

Human activities have increased with urbanisation in the Erhai Lake Basin, considerably impacting its eco-environmental quality (EEQ). This study aims to reveal the evolution and driving forces of the EEQ using water benefit-based ecological index (WBEI) in response to human activities and policy variations in the Erhai Lake Basin from 1990 to 2020. Results show that (1) the EEQ exhibited a pattern of initial degradation, subsequent improvement, further degradation and a rebound from 1990 to 2020, and the areas with poor and fair EEQ levels mainly concentrated around the Erhai Lake Basin with a high level of urbanisation and relatively flat terrain; (2) the EEQ levels were not optimistic in 1990, 1995 and 2015, and areas with poor and fair EEQ levels accounted for 43.41%, 47.01% and 40.05% of the total area, respectively; and (3) an overall improvement in the EEQ was observed in 1995-2000, 2000-2005, 2005-2009 and 2015-2020, and the improvement was most significant in 1995-2000, covering an area of 823.95 km2 and accounting for 31.79% of the total area. Results also confirmed that the EEQ changes in the Erhai Lake Basin were primarily influenced by human activities and policy variations. Moreover, these results can provide a scientific basis for the formulation and planning of sustainable development policy in the Erhai Lake Basin.

Correlations between ecosystem service value and landscape ecological risk and its spatial heterogeneity in Jilin Province, China.

Exploring the correlations between ecosystem service value (ESV) and landscape ecological risk and the driving factors of their spatial variations is crucial for maintaining regional ecological security and promoting sustainable human well-being. We carried out a grid resampling size of 5 km×5 km assessment units of Jilin Pro-vince based on the remote sensing monitoring data of land use in 2000, 2005, 2010, 2015, and 2020. We quantitatively evaluated the landscape ecological risk and ESV, and analyzed their spatial-temporal variations. Employing bivariate spatial autocorrelation analysis and the geographical detector models, we examined the correlation between the landscape ecological risk and ESV and explored the driving factors for their spatial variations. The results showed that ESV in Jilin Province decreased from 385.895 billion yuan to 378.211 billion yuan during 2000-2020. The eastern region was dominated by extremely low risk, medium risk, and low risk areas. In contrast, the western region was mainly composed of extremely high risk and high risk areas. There was a significant negative correlation and spatial negative correlation between landscape ecological risk and ESV in Jilin Province. Human activity and land use type were the important driving factors for spatial differentiation in both landscape ecological risk and ESV. Our findings suggested that scientific land use regulation and appropriate control of human activities are critically needed to optimize Jilin Province's ecological environment.

HARNet in deep learning approach-a systematic survey.

A comprehensive examination of human action recognition (HAR) methodologies situated at the convergence of deep learning and computer vision is the subject of this article. We examine the progression from handcrafted feature-based approaches to end-to-end learning, with a particular focus on the significance of large-scale datasets. By classifying research paradigms, such as temporal modelling and spatial features, our proposed taxonomy illuminates the merits and drawbacks of each. We specifically present HARNet, an architecture for Multi-Model Deep Learning that integrates recurrent and convolutional neural networks while utilizing attention mechanisms to improve accuracy and robustness. The VideoMAE v2 method ( https://github.com/OpenGVLab/VideoMAEv2 ) has been utilized as a case study to illustrate practical implementations and obstacles. For researchers and practitioners interested in gaining a comprehensive understanding of the most recent advancements in HAR as they relate to computer vision and deep learning, this survey is an invaluable resource.

Adversarial AI applied to cross-user inter-domain and intra-domain adaptation in human activity recognition using wireless signals.

In recent years, researchers have successfully recognised human activities using commercially available WiFi (Wireless Fidelity) devices. The channel state information (CSI) can be gathered at the access point with the help of a network interface controller (NIC card). These CSI streams are sensitive to human body motions and produce abrupt changes (fluctuations) in their magnitude and phase values when a moving object interacts with a transmitter and receiver pair. This sensing methodology is gaining popularity compared to traditional approaches involving wearable technology, as it is a contactless sensing strategy with no cumbersome sensing equipments fitted on the target with preserved privacy since no personal information of the subject is collected. In previous investigations, internal validation statistics have been promising. However, external validation results have been poor, due to model application to varying subjects with remarkably different environments. To address this problem, we propose an adversarial Artificial Intelligence AI model that learns and utilises domain-invariant features. We analyse model results in terms of suitability for inter-domain and intra-domain alignment techniques, to identify which is better at robustly matching the source to target domain, and hence improve recognition accuracy in cross-user conditions for HAR using wireless signals. We evaluate our model performance on different target training data percentages to assess model reliability on data scarcity. After extensive evaluation, our architecture shows improved predictive performance across target training data proportions when compared to a non-adversarial model for nine cross-user conditions with comparatively less simulation time. We conclude that inter-domain alignment is preferable for HAR applications using wireless signals, and confirm that the dataset used is suitable for investigations of this type. Our architecture can form the basis of future studies using other datasets and/or investigating combined cross-environmental and cross-user features.

Comparative evaluation of soundscapes in human activities spatial contexts of pedestrian spaces adjacent to arterial roads.

Pedestrian spaces adjacent to arterial roads are characterized by the dominance of traffic noise alongside various human activities. Research on the impact of traffic noise on the soundscape evaluation of pedestrian spaces has not considered human activities spatial contexts. To address this research gap, the present study constructed auditory environments for pedestrian spaces in the contexts of commuting, residential, and commercial activities. A total of seven auditory environments were subjected to laboratory auditory evaluations, including perceived dominance of sound source, acoustic comfort, and perceived affective quality of the soundscape. The results indicated that in pedestrian spaces with constant traffic noise, the presence of significant human activity sounds led to a decreased perceived dominance of traffic noise and an increased acoustic comfort, despite the higher acoustic energy. Thus, pedestrian spaces with a variety of human activity received better soundscape evaluations. The elements that reflected the human activities spatial contexts, including the types and intensity of human activities, played a crucial role in soundscape evaluations. Better acoustic comfort was reported in pedestrian spaces characterized by low-intensity residential activities and high-intensity commercial activities. Additionally, pedestrian spaces with more intense activities offered an actively engaging soundscape. The findings can provide reference for a more accurate evaluation of the soundscape in pedestrian spaces and guide the soundscape design of pedestrian environments.

Quantitative Analysis of Mother Wavelet Function Selection for Wearable Sensors-Based Human Activity Recognition.

Recent advancements in the Internet of Things (IoT) wearable devices such as wearable inertial sensors have increased the demand for precise human activity recognition (HAR) with minimal computational resources. The wavelet transform, which offers excellent time-frequency localization characteristics, is well suited for HAR recognition systems. Selecting a mother wavelet function in wavelet analysis is critical, as optimal selection improves the recognition performance. The activity time signals data have different periodic patterns that can discriminate activities from each other. Therefore, selecting a mother wavelet function that closely resembles the shape of the recognized activity's sensor (inertial) signals significantly impacts recognition performance. This study uses an optimal mother wavelet selection method that combines wavelet packet transform with the energy-to-Shannon-entropy ratio and two classification algorithms: decision tree (DT) and support vector machines (SVM). We examined six different mother wavelet families with different numbers of vanishing points. Our experiments were performed on eight publicly available ADL datasets: MHEALTH, WISDM Activity Prediction, HARTH, HARsense, DaLiAc, PAMAP2, REALDISP, and HAR70+. The analysis demonstrated in this paper can be used as a guideline for optimal mother wavelet selection for human activity recognition.

Soil quality assessment in low human activity disturbance zones: a study on the Qinghai-Tibet Plateau.

The Qinghai-Tibet Plateau, located at the Third Pole and known as the "Asian water tower," serves as a crucial ecological barrier for China. Grasping the soil quality on the Qinghai-Tibet Plateau holds paramount importance for the rational and scientific exploitation of soil resources within the region and is essential for vegetation restoration and ecological reconstruction. This study, conducted in Maqin County, Qinghai Province, collected 1647 soil samples (0-20 cm) within a study area of 6300 km2. Sixteen soil indicators were selected that were split into beneficial (N, P, S, and B), harmful (Cr, Hg, As, Pb, Ni, and Cd), and essential (Cu, Zn, Se, Ga, K, and Ca) elements. The Soil Quality Index (SQI) was computed to assess soil quality across diverse geological contexts, land cover classifications, and soil profiles. The results indicate that the overall SQI in the study area was comparatively high, with most regions having an SQI between 0.4 and 0.6, categorized as moderately to highly satisfactory. Among the different geological backgrounds, the highest SQI was found in the Quaternary alluvium (0.555) and the lowest in the Precambrian Jinshuikou Formation (0.481). Regarding different land-use types, the highest SQI was observed in glacier- and snow-covered areas (0.582) and the lowest in other types of grassland (0.461). The highest SQI was recorded in typical alpine meadow soil (0.521) and the lowest in leached brown soil (0.460). The evaluation results have significant reference value for the sustainable utilization and management of soil in Maqin County, Qinghai Province, China.

A Multi-Modal Egocentric Activity Recognition Approach towards Video Domain Generalization.

Egocentric activity recognition is a prominent computer vision task that is based on the use of wearable cameras. Since egocentric videos are captured through the perspective of the person wearing the camera, her/his body motions severely complicate the video content, imposing several challenges. In this work we propose a novel approach for domain-generalized egocentric human activity recognition. Typical approaches use a large amount of training data, aiming to cover all possible variants of each action. Moreover, several recent approaches have attempted to handle discrepancies between domains with a variety of costly and mostly unsupervised domain adaptation methods. In our approach we show that through simple manipulation of available source domain data and with minor involvement from the target domain, we are able to produce robust models, able to adequately predict human activity in egocentric video sequences. To this end, we introduce a novel three-stream deep neural network architecture combining elements of vision transformers and residual neural networks which are trained using multi-modal data. We evaluate the proposed approach using a challenging, egocentric video dataset and demonstrate its superiority over recent, state-of-the-art research works.

Human Action Recognition and Note Recognition: A Deep Learning Approach Using STA-GCN.

Human action recognition (HAR) is growing in machine learning with a wide range of applications. One challenging aspect of HAR is recognizing human actions while playing music, further complicated by the need to recognize the musical notes being played. This paper proposes a deep learning-based method for simultaneous HAR and musical note recognition in music performances. We conducted experiments on Morin khuur performances, a traditional Mongolian instrument. The proposed method consists of two stages. First, we created a new dataset of Morin khuur performances. We used motion capture systems and depth sensors to collect data that includes hand keypoints, instrument segmentation information, and detailed movement information. We then analyzed RGB images, depth images, and motion data to determine which type of data provides the most valuable features for recognizing actions and notes in music performances. The second stage utilizes a Spatial Temporal Attention Graph Convolutional Network (STA-GCN) to recognize musical notes as continuous gestures. The STA-GCN model is designed to learn the relationships between hand keypoints and instrument segmentation information, which are crucial for accurate recognition. Evaluation on our dataset demonstrates that our model outperforms the traditional ST-GCN model, achieving an accuracy of 81.4%.

Human Activity Recognition Based on Deep Learning and Micro-Doppler Radar Data.

Activity recognition is one of the significant technologies accompanying the development of the Internet of Things (IoT). It can help in recording daily life activities or reporting emergencies, thus improving the user's quality of life and safety, and even easing the workload of caregivers. This study proposes a human activity recognition (HAR) system based on activity data obtained via the micro-Doppler effect, combining a two-stream one-dimensional convolutional neural network (1D-CNN) with a bidirectional gated recurrent unit (BiGRU). Initially, radar sensor data are used to generate information related to time and frequency responses using short-time Fourier transform (STFT). Subsequently, the magnitudes and phase values are calculated and fed into the 1D-CNN and Bi-GRU models to extract spatial and temporal features for subsequent model training and activity recognition. Additionally, we propose a simple cross-channel operation (CCO) to facilitate the exchange of magnitude and phase features between parallel convolutional layers. An open dataset collected through radar, named Rad-HAR, is employed for model training and performance evaluation. Experimental results demonstrate that the proposed 1D-CNN+CCO-BiGRU model demonstrated superior performance, achieving an impressive accuracy rate of 98.2%. This outperformance of existing systems with the radar sensor underscores the proposed model's potential applicability in real-world scenarios, marking a significant advancement in the field of HAR within the IoT framework.

Adopting Graph Neural Networks to Analyze Human-Object Interactions for Inferring Activities of Daily Living.

Human Activity Recognition (HAR) refers to a field that aims to identify human activities by adopting multiple techniques. In this field, different applications, such as smart homes and assistive robots, are introduced to support individuals in their Activities of Daily Living (ADL) by analyzing data collected from various sensors. Apart from wearable sensors, the adoption of camera frames to analyze and classify ADL has emerged as a promising trend for achieving the identification and classification of ADL. To accomplish this, the existing approaches typically rely on object classification with pose estimation using the image frames collected from cameras. Given the existence of inherent correlations between human-object interactions and ADL, further efforts are often needed to leverage these correlations for more effective and well justified decisions. To this end, this work proposes a framework where Graph Neural Networks (GNN) are adopted to explicitly analyze human-object interactions for more effectively recognizing daily activities. By automatically encoding the correlations among various interactions detected through some collected relational data, the framework infers the existence of different activities alongside their corresponding environmental objects. As a case study, we use the Toyota Smart Home dataset to evaluate the proposed framework. Compared with conventional feed-forward neural networks, the results demonstrate significantly superior performance in identifying ADL, allowing for the classification of different daily activities with an accuracy of 0.88. Furthermore, the incorporation of encoded information from relational data enhances object-inference performance compared to the GNN without joint prediction, increasing accuracy from 0.71 to 0.77.