Abnormal Behavior Research Articles

A method for recognizing abnormal behaviors of personnel at petroleum stations based on GTB-ResNet

PurposeThe abnormal behaviors of staff at petroleum stations pose significant safety hazards. Addressing the challenges of high parameter counts, lengthy training periods and low recognition rates in existing 3D ResNet behavior recognition models, this paper proposes GTB-ResNet, a network designed to detect abnormal behaviors in petroleum station staff.Design/methodology/approachFirstly, to mitigate the issues of excessive parameters and computational complexity in 3D ResNet, a lightweight residual convolution module called the Ghost residual module (GhostNet) is introduced in the feature extraction network. Ghost convolution replaces standard convolution, reducing model parameters while preserving multi-scale feature extraction capabilities. Secondly, to enhance the model's focus on salient features amidst wide surveillance ranges and small target objects, the triplet attention mechanism module is integrated to facilitate spatial and channel information interaction. Lastly, to address the challenge of short time-series features leading to misjudgments in similar actions, a bidirectional gated recurrent network is added to the feature extraction backbone network. This ensures the extraction of key long time-series features, thereby improving feature extraction accuracy.FindingsThe experimental setup encompasses four behavior types: illegal phone answering, smoking, falling (abnormal) and touching the face (normal), comprising a total of 892 videos. Experimental results showcase GTB-ResNet achieving a recognition accuracy of 96.7% with a model parameter count of 4.46 M and a computational complexity of 3.898 G. This represents a 4.4% improvement over 3D ResNet, with reductions of 90.4% in parameters and 61.5% in computational complexity.Originality/valueSpecifically designed for edge devices in oil stations, the 3D ResNet network is tailored for real-time action prediction. To address the challenges posed by the large number of parameters in 3D ResNet networks and the difficulties in deployment on edge devices, a lightweight residual module based on ghost convolution is developed. Additionally, to tackle the issue of low detection accuracy of behaviors amidst the noisy environment of petroleum stations, a triple attention mechanism is introduced during feature extraction to enhance focus on salient features. Moreover, to overcome the potential for misjudgments arising from the similarity of actions, a Bi-GRU model is introduced to enhance the extraction of key long-term features.

The Acute and Subchronic Toxicity Evaluation of methanol leaf Extracts Of Bombaxbuonopozense

Natural products serve as the foundation for conventional medicine and are significant sources of lead compounds for medication research. One of the underutilized plants in West Africa is Bombax buonopozense. West African populations use the majority of its parts for nutritional and therapeutic purposes. The research was aimed at evaluation of the Acute and subchronic oral toxicity of Bombaxbuonopozense leaves extract in rats.In the acute toxicity study, female rats received oral doses of 2000 mg/kg of Bombaxbuonopozense extract (n = 5/group). Abnormal behavior, toxic symptoms, weight changes, and mortality were monitored for 14 consecutive days to evaluate acute toxicity. For the sub-chronic toxicity study, the extract was orally administered to Wistar rats at doses of 800 and 1600 mg/kg (n = 5/group) daily for 28 days. Daily observations were made on the general behavior and body weight of the rats. Biochemical and hematological analyses were performed at the end of the treatment period. Following acute oral administration, the LD50 of Bombaxbuonopozens leaves extract was found to be greater than 2000 mg/kg in the evaluation of acute oral toxicity. Administration of Bombaxbuonopozense extract to rats did showed significantly increase in platelate in the treated group (362.00±23.67*) when compared to the control group (451.67±30.77) at (p>0.05). The liver enzymes, An increase (P>0.05) was seen in ALP concentration on the treated group (216.59±44.29*) In the subchronic oral toxicity assessment at doses of 800 and 1600 mg/kg/day for 28 days, no significant changes were observed in food consumption, body weight gain, organ weights, and biochemical parameters. Treatment of rats with Bombaxbuonopozense extract showed a significant increase in ALP on the group treated with 1600 mg/kg (216.59±44.29*) compared to the control (98.70±13.32) and significant increase in LDL for the treated groups 800 mg/kg (70.76±8.19*) and 1600 mg/kg (71.40±13.17*) were observed when compared with.The current research demonstrated that administration of Bombaxbuonopozense extract orally for 28 days, at doses of up to 1600 mg/kg, did not cause any toxic effects. Based on the acute toxicity study, the estimated median lethal dose (LD50) of the extract was more than 2000 mg/kg.

Improved identification of network anomalies through optimal CURE clustering

Abstract In this paper, we propose an advanced network anomaly behavior identification framework to overcome the constraints inherent in conventional rule- or signature-based approaches, which often struggle with emerging and previously unknown threats. Central to our framework is an Enhanced CURE (Cluster Updating and REfining) clustering algorithm, meticulously tailored and refined to incorporate a density-based methodology. This enhancement enables the algorithm to discern subtle shifts in network anomaly patterns with heightened precision. The implementation workflow commences with the application of the optimized CURE algorithm to analyze network data, followed by the deployment of a sophisticated anomaly degree ranking mechanism. This mechanism, through meticulous calculation of individual data points’ anomaly degrees and subsequent ranking, effectively isolates those deviating significantly from standard behavioral norms, incorporating a strategic threshold to filter out false positives. To validate our methodology’s efficacy and its superiority over existing techniques, experiments were conducted utilizing a substantial real-world network dataset. These tests affirm not only a marked increase in the accuracy of abnormal behavior recognition and a reduction in computational intricacy but also demonstrate the adaptability across diverse network ecosystems. Our approach has proven successful in pinpointing a wide array of network anomalies, encompassing malicious cyberattacks, fraudulent activities, unauthorized intrusions, and breaches of security protocols, thereby highlighting its comprehensive capability in bolstering network defense strategies. Despite the notable advancements and successful identification of various network anomalies, our framework currently lacks integration with real-time learning capabilities, limiting its immediate responsiveness to rapidly evolving attack patterns and necessitating ongoing research for dynamic updates and adaptive learning mechanisms.

Chronic Sleep Deprivation Altered the Expression of Memory-Related Genes and Caused Cognitive Memory Dysfunction in Mice.

Lack of sleep, whether acute or chronic, is quite common and negatively affects an individual's memory and cognitive function. The question of whether chronic sleep deprivation (CSD) causes cognitive impairment to arise and progress is not well studied. To investigate the effects of CSD on memory and cognition, this study began by establishing a CSD mouse model. Behavioral experiments on animals revealed that CSD induced cognitive behavioral abnormalities reminiscent of Alzheimer's disease. Western blot experiments further demonstrated a considerable increase in amyloid-β (Aβ) expression in the mouse brain following CSD. Meanwhile, the hub gene Prkcg was searched for in the cerebellum using RNA-seq and bioinformatics analysis. PKCγ (Prkcg) expression was significantly reduced, as demonstrated by RT-qPCR and Western blot validations. Additionally, CSD was associated with downregulated CREB expression, decreased expression of the endothelin-converting enzyme (ECE1), and increased phosphorylation of ERK1/2 downstream of PKCγ. These findings suggested that CSD down-regulated PKCγ expression, decreased ECE1 expression, impaired Aβ degradation, and affected the PKCγ/ERK/CREB pathway and the synthesis of memory-related proteins. Overall, this study highlighted how CSD modulated PKCγ-related metabolism, impacting Aβ clearance and the production of memory-related proteins. Such insights are crucial for understanding and preventing sporadic Alzheimer's disease (sAD) associated with CSD.

Application of data-driven Decision Support System in Combating Illegal Wildlife Trade on Internet

With the rapid development of the Internet, the illegal wildlife trade is becoming increasingly serious, posing a threat to biodiversity. Using big data and artificial intelligence technology, this paper identifies abnormal transactions and network relationships by analyzing social media and e-commerce platforms, combined with random forests and graph neural networks, to enhance the efficiency of the crackdown. Random forest algorithms reveal nonlinear patterns and abnormal behavior, and graph neural networks help reveal the complex structure of the transaction network. In implementing the strategy, this paper uses a multi-attribute decision-making model (MADM) and analytic hierarchy Process (AHP) to assess and select the best partners, such as the United Nations Environment Programme, to optimize resource allocation and synergies. Through these technologies, this paper hopes to significantly increase the precision and speed of combating illegal wildlife trade, thereby effectively protecting biodiversity and contributing to the sustainable development of the planet and the harmonious coexistence of the natural environment.

Abnormal Temperature Dependence of Huang-Rhys Factor and Exciton Recombination Kinetics in CsPbBr3 Perovskite Quantum Dots.

Anomalous thermal behaviors of excitonic luminescence in CsPbBr3 perovskite quantum dots (PQDs) were observed. It is found that the main luminescence peak originated from the excitonic radiative recombination assisted by the longitudinal-optical (LO) phonon, and its integrated intensity first declines as the temperature varies from 10 to 150 K and then turns to increase at ∼160 K, reaching a maximum value at 300 K. A model considering the thermal detrapping and transfer of electrons from a trap level is developed to interpret these abnormal thermal behaviors of the luminescence from the PQDs. On the other hand, the quantum-mechanical multimode Brownian oscillator model was employed to unravel that the Huang-Rhys factor exclusively characterizing the exciton-phonon coupling in the studied CsPbBr3 PQDs decreases from 1.65 at 10 K to 1.31 at 200 K.

Distributed Photovoltaic Communication Anomaly Detection Based on Spatiotemporal Feature Collaborative Modeling

As distributed photovoltaic (PV) technology rapidly develops and is widely applied, the methods of cyberattacks are continuously evolving, posing increasingly severe threats to the communication networks of distributed PV systems. Recent studies have shown that the Transformer model, which effectively integrates global information and handles long-distance dependencies, has garnered significant attention. Based on this, our research proposes a model named STformer, which is applied to the task of attack detection in distributed PV communication. Specifically, we propose a temporal attention mechanism and a variable attention mechanism. The temporal attention mechanism focuses on capturing subtle changes and trends in data sequences over time, ensuring a highly sensitive recognition of patterns inherent in time-series data. In contrast, the variable attention mechanism analyzes the intrinsic relationships and interactions between different variables, uncovering critical correlations that may indicate abnormal behavior or potential attacks. Additionally, we incorporate the Uniform Manifold Approximation and Projection (UMAP) dimensionality reduction technique. This technique not only helps reduce computational complexity but, in certain cases, can enhance anomaly detection performance. Finally, compared to classical and advanced methods, STformer demonstrates satisfactory performance in simulation experiments.

Motor, mood, and memory impairments persist during remission periods in chronic colitis and are influenced by neuroinflammation and sex.

Ulcerative colitis is a chronic pathology characterized by relapsing-remitting phases of intestinal inflammation. Additionally, some patients develop neuropsychiatric disorders, such as depression and anxiety, or cognitive deficits. We aimed to investigate whether the development of chronic colitis elicits memory, locomotion, and mood impairments. It further examined whether these impairments are influenced by the relapsing-remitting phases of the colitis or by sex. Here, we used a chronic colitis model in male and female rats, induced with sodium dextran sulfate, mirroring the phases of human ulcerative colitis. Our results revealed that the severity of colitis was slightly higher in males than females. Chronic colitis triggered motor and short-term memory deficits and induced anxiety- and depression-like behaviors that remained throughout the development of the disease. There are also sex differences under control or inflammatory conditions. Therefore, in both situations, females compared to males displayed: (i) slightly lower locomotion, (ii) increased anxiety-like behaviors, (iii) similar depression-like behaviors, and (iv) similar short-term memory deficit. Additionally, under control conditions, the mRNA levels of IL-1β, IL-6, and TNF-α were higher in the female hippocampus. In both sexes, when chronic colitis was established, the neuroinflammation was evidenced by increased mRNA levels of these three cytokines in the hippocampus and in the motor and prefrontal cortices. Interestingly, this neuroinflammation was slightly greater in males. In summary, we show that the development of chronic colitis caused persistent behavioral abnormalities, highlighting sex differences, and that could be a consequence, at least in part, of the increase in IL-1β, IL-6, and TNF-α in the brain.

A Review of Abnormal Crowd Behavior Recognition Technology Based on Computer Vision

Abnormal crowd behavior recognition is one of the research hotspots in computer vision. Its goal is to use computer vision technology and abnormal behavior detection models to accurately perceive, predict, and intervene in potential abnormal behaviors of the crowd and monitor the status of the crowd system in public places in real time, to effectively prevent and deal with public security risks and ensure public life safety and social order. To this end, focusing on the abnormal crowd behavior recognition technology in the computer vision system, a systematic review study of its theory and cutting-edge technology is conducted. First, the crowd level and abnormal behaviors in public places are defined, and the challenges faced by abnormal crowd behavior recognition are expounded. Then, from the dimensions based on traditional methods and based on deep learning, the mainstream technologies of abnormal behavior recognition are discussed, and the design ideas, advantages, and limitations of various methods are analyzed. Next, the mainstream software tools are introduced to provide a comprehensive reference for the technical framework. Secondly, typical abnormal behavior datasets at home and abroad are sorted out, and the characteristics of these datasets are compared in detail from multiple perspectives such as scale, characteristics, and uses, and the performance indicators of different algorithms on the datasets are compared and analyzed. Finally, the full text is summarized and the future development direction of abnormal crowd behavior recognition technology is prospected.

Associations between genotype, phenotype and behaviours measured by the Rett syndrome behaviour questionnaire in Rett syndrome

IntroductionRett syndrome (RTT) is a rare neurodevelopmental disorder with developmental impairments, comorbidities, and abnormal behaviours such as hand stereotypies and emotional features. The Rett Syndrome Behaviour Questionnaire (RSBQ) was developed to describe the behavioural and emotional features of RTT. Little is known how RSBQ scores are associated with genetic and clinical characteristics in RTT. This study investigated relationships between genotype, age, walking, hand function, sleep, and RSBQ total and subscale scores in RTT.MethodsThis is a cross-sectional analysis of data collected in the Australian Rett Syndrome Database and the International Rett Syndrome Phenotype Database. Parent caregivers completed the RSBQ and Sleep Disturbance Scale for Children [subscales for disorders of initiating and maintaining sleep (DIMS), disorders of excessive somnolence (DOES)], and provided information on age, variant type, functional abilities (mobility, hand function), seizure frequency and gastrointestinal problems. Associations between the RSBQ scores and the independent variables were modelled using linear regression.ResultsData were available for 365 individuals with RTT [median (range) age 17.8 (2.9–51.9) years, 2 males]. Compared to adults, 2- to 12-year-old children had higher mean Total, Night-time Behaviour and Fear/Anxiety scores. Compared to individuals with a C-terminal deletion, individuals with the p.Arg255* variant had higher mean Total and Night-time Behaviours scores, whereas the p.Arg294* variant had higher mean Mood scores. Individuals with intermediate mobility and hand function abilities had a higher mean Total score. Total RSBQ and subscale scores were similar across categories for seizures, constipation, and reflux, but were higher with abnormal DIMS and abnormal DOES scores.ConclusionExcept for associations with sleep, the RSBQ measures the behavioural phenotype rather than clinical severity in RTT, as traditionally conceptualised in terms of functional abilities and comorbidities. When designing clinical trials, the RSBQ needs to be complemented by other outcome measures to assess specific core functions and associated comorbidities in RTT.

A Flexible Smart Healthcare Platform Conjugated with Artificial Epidermis Assembled by Three-Dimensionally Conductive MOF Network for Gas and Pressure Sensing

The rising flexible and intelligent electronics greatly facilitate the noninvasive and timely tracking of physiological information in telemedicine healthcare. Meticulously building bionic-sensitive moieties is vital for designing efficient electronic skin with advanced cognitive functionalities to pluralistically capture external stimuli. However, realistic mimesis, both in the skin’s three-dimensional interlocked hierarchical structures and synchronous encoding multistimuli information capacities, remains a challenging yet vital need for simplifying the design of flexible logic circuits. Herein, we construct an artificial epidermal device by in situ growing Cu3(HHTP)2 particles onto the hollow spherical Ti3C2Tx surface, aiming to concurrently emulate the spinous and granular layers of the skin’s epidermis. The bionic Ti3C2Tx@Cu3(HHTP)2 exhibits independent NO2 and pressure response, as well as novel functionalities such as acoustic signature perception and Morse code-encrypted message communication. Ultimately, a wearable alarming system with a mobile application terminal is self-developed by integrating the bimodular senor into flexible printed circuits. This system can assess risk factors related with asthmatic, such as stimulation of external NO2 gas, abnormal expiratory behavior and exertion degrees of fingers, achieving a recognition accuracy of 97.6% as assisted by a machine learning algorithm. Our work provides a feasible routine to develop intelligent multifunctional healthcare equipment for burgeoning transformative telemedicine diagnosis.

An intelligent recognition method of factory personnel behavior based on deep learning

The real-time and accurate recognition of abnormal behavior among factory personnel helps enhance their awareness of hazardous environments, thereby reducing the occurrence of accidents. This paper proposes a behavior recognition network based on an attention mechanism and a high-efficiency convolution module. The Bi-Level Routing Attention was introduced to the backbone network, thus enhancing the attention of the recognition network to the target region effectively. The recognition accuracy was further strengthened by improving the neck network based on the ConvNeXt Block module while reducing the model complexity. Thirteen additional recognition models were constructed to enhance the original network from various perspectives. Subsequently, the mean average precision and detection speed of each model were evaluated. Experimental results demonstrated that the detection accuracy of the target recognition network proposed in this paper has been significantly improved, the detection speed meets the real-time requirements, and the comprehensive performance is the most superior compared with other diverse and improved networks. The proposed recognition model can accurately identify a variety of factory personnel's abnormal behaviors in real-time, and it has practical application significance for the problem of personnel safety identification in the factory.

BCL11A intellectual developmental disorder: defining the clinical spectrum and genotype-phenotype correlations.

An increasing number of individuals with intellectual developmental disorder (IDD) and heterozygous variants in BCL11A are identified, yet our knowledge of manifestations and mutational spectrum is lacking. To address this, we performed detailed analysis of 42 individuals with BCL11A-related IDD (BCL11A-IDD, a.k.a. Dias-Logan syndrome) ascertained through an international collaborative network, and reviewed 35 additional previously reported patients. Analysis of 77 affected individuals identified 60 unique disease-causing variants (30 frameshift, 7 missense, 6 splice-site, 17 stop-gain) and 8 unique BCL11A microdeletions. We define the most prevalent features of BCL11A-IDD: IDD, postnatal-onset microcephaly, hypotonia, behavioral abnormalities, autism spectrum disorder, and persistence of fetal hemoglobin (HbF), and identify autonomic dysregulation as new feature. BCL11A-IDD is distinguished from 2p16 microdeletion syndrome, which has a higher incidence of congenital anomalies. Our results underscore BCL11A as an important transcription factor in human hindbrain development, identifying a previously underrecognized phenotype of a small brainstem with a reduced pons/medulla ratio. Genotype-phenotype correlation revealed an isoform-dependent trend in severity of truncating variants: those affecting all isoforms are associated with higher frequency of hypotonia, and those affecting the long (BCL11A-L) and extra-long (-XL) isoforms, sparing the short (-S), are associated with higher frequency of postnatal microcephaly. With the largest international cohort to date, this study highlights persistence of fetal hemoglobin as a consistent biomarker and hindbrain abnormalities as a common feature. It contributes significantly to our understanding of BCL11A-IDD through an extensive unbiased multi-center assessment, providing valuable insights for diagnosis, management and counselling, and into BCL11A's role in brain development.

No effect of infant nursery rearing on laboratory rhesus monkey dams’ social behavior or long-term cortisol profiles

While the behavioral and physiological impacts of nursery rearing in laboratory-housed infant monkeys have been well characterized, to date no studies have examined the impact on nonhuman primate dams of their infants being removed for rearing in the nursery. Despite the lack of evidence on the topic, anti-animal research groups often cite mother-infant separation and infant nursery rearing (NR) as a welfare concern for laboratory monkey mothers. As such, important policy decisions regarding research activities may result without adequate evidence. Therefore, we designed this study to examine behavioral and hormonal responses in laboratory monkey dams, who were part of independent long-term longitudinal studies, in response to their infants being NR or mother-peer-reared (MPR). We assessed social, self-care, environmental exploration, and abnormal behaviors for 30 days before and 30 days after parturition in rhesus monkey dams (Macaca mulatta, N=49). Infants were randomly assigned to be either NR (N = 27) or MPR (N = 22). We also analyzed hair samples for cortisol concentrations three times across the study period (Pregnancy, Neonatal Period, Peak Lactation). Dams of NR infants showed no gross behavioral differences relative to dams of MPR infants (all p’s>0.05). Dams of MPR infants showed expected increases in social grooming and social contact, and concomitant decreases in foraging and locomotion, in the 30 days post-partum compared to the 30 days pre-partum (p<0.01). Dams whose infants were NR or MPR showed no differences in hair cortisol concentrations across the study period (p>0.05). We conclude that, with respect to the behavioral and endocrine measures we assessed, nursery rearing of their infants is not detrimental to the welfare of laboratory-housed macaque dams.

Construction and Application of a Private 5G Standalone Medical Network in a Smart Health Environment: Exploratory Practice From China.

To date, the differentiated requirements for network performance in various health care service scenarios-within, outside, and between hospitals-remain a key challenge that restricts the development and implementation of digital medical services. This study aims to construct and implement a private 5G (the 5th generation mobile communication technology) standalone (SA) medical network in a smart health environment to meet the diverse needs of various medical services. Based on an analysis of network differentiation requirements in medical applications, the system architecture and functional positioning of the proposed private 5G SA medical network are designed and implemented. The system architecture includes the development of exclusive and preferential channels for medical use, as well as an ordinary user channel. A 3-layer network function architecture is designed, encompassing resource, control, and intelligent operation layers to facilitate management arrangements and provide network open services. Core technologies, including edge cloud collaboration; service awareness; and slicing of access, bearer, and core networks, are employed in the construction and application of the 5G SA network. The construction of the private 5G SA medical network primarily involves system architecture, standards, and security measures. The system, featuring exclusive, preferential, and common channels, supports a variety of medical applications. Relevant standards are adhered to in order to ensure the interaction and sharing of medical service information. Security is achieved through mechanisms such as authentication, abnormal behavior analysis, and dynamic access control. Three typical medical applications that rely on the 5G network in intrahospital, interhospital, and out-of-hospital scenarios-namely, mobile ward rounds, remote first aid, and remote ultrasound-were conducted. Testing of the 5G-enabled mobile ward rounds showed an average download rate of 790 Mbps and an average upload rate of 91 Mbps. Compared with 4G, the 5G network more effectively meets the diverse requirements of various business applications in prehospital emergency scenarios. For remote ultrasound, the average downlink rate of the 5G network is 4.82 Mbps, and the average uplink rate is 2 Mbps, with an average fluctuation of approximately 8 ms. The bandwidth, performance, and delay of the 5G SA network were also examined and confirmed to be effective. The proposed 5G SA medical network demonstrates strong performance in typical medical applications. Its construction and application could lead to the development of new medical service models and provide valuable references for the further advancement and implementation of 5G networks in other industries, both in China and globally.

STAD: Ship trajectory anomaly detection in ocean with dynamic pattern clustering

With the continuous increase of maritime traffic, developing an efficient and accurate model for ship trajectory anomaly detection has become crucial for ensuring maritime transportation safety. The high complexity and variability of the marine environment lead to diverse ship trajectory patterns, making it challenging to learn effective trajectory representations for accurately identifying anomalies. We thus proposed an unsupervised deep learning model called STAD for ship trajectory anomaly detection in ocean to address this challenge. Concretely, STAD leverages offset reconstruction-based representation learning and a deep Gaussian Mixture Model (GMM) estimation network to learn the underlying complex clustering patterns of ship trajectories and utilize the learned patterns to enhance trajectory anomaly detection. Extensive experiments on multiple AIS datasets indicate that our model significantly outperforms existing methods in detecting multiple representative types of ship trajectory anomalies, including shift deviation, abnormal heading, and abnormal speeding. This study could help closely monitor the status of ship movement and detect abnormal behaviors in advance, thus benefiting maritime safety.

Research on bridge safety early warning method based on strain energy theory and health monitoring data

Bridges are technology-intensive and heavily invested permanent infrastructure. After completion and opening to traffic, bridge structures are easy to be affected by factors such as traffic load and atmospheric environment. Therefore, it is necessary to do safety warning and evaluation of bridges, especially the abnormal behavior in the early stages of bridge operation. In this research paper, a large-span continuous rigid frame bridge installed with the health monitoring system (HMS), of which a large amount of health monitoring data are collected by the HMS, is used as an example, and then a bridge safety early warning method is proposed when the bridge is during early operating period. First of all, the research finding that the internal stress of the prototype bridge obeys normal distribution through statistical analysis is used; next, we deduce that the strain energy inside the prototype bridge is subject to the Non-central Chi-square distribution combined with the strain energy and statistical theories; in the end, the key probability density distribution function of strain energy and its parameters are derived by using the key stress distribution function of the high performance concrete C50 strength grade used in the prototype bridge. The method recommended in this paper is conducive to the formulation of bridge preventive maintenance strategies.

A Novel WTG Method for Predicting Ship Trajectories in the Fujian Inshore Area Based on AIS Data

The increasing congestion in major global maritime routes poses significant threats to international maritime safety, exacerbated by the proliferation of large, high-speed vessels. To improve the detection of abnormal ship behavior, this research employed automatic identification system (AIS) data for ship trajectory forecasting. Traditional methods primarily focus on spatial and temporal correlations but often lack accuracy and reliability. In this study, ship path predictions were enhanced using the WTG model, which combines wavelet transform, temporal convolutional networks (TCN), and gated recurrent units (GRU). Initially, wavelet decomposition was applied to deconstruct the input trajectory time series. The TCN and GRU modules then extracted features from both the time series and the decomposed data. The predicted elements were reassembled using a multi-head attention mechanism and a pooling layer to produce the final predictions. Comparative experiments demonstrated that the WTG model surpasses other models in the accuracy of ship trajectory prediction. The model proposed in this study proves to be reliable for forecasting ship paths, which is crucial for marine traffic management and ensuring safe navigation.

CSSDet: small object detection via cross-scale feature enhancement on drone-view images

ABSTRACT Object detection on drone-view images is vital for applications like intelligent transportation, abnormal behavior detection, and urban surveillance. However, the diverse perspectives and altitudes from which Unmanned Aerial Vehicle (UAV) capture scenes pose challenges due to multi-scale differences and small object sizes. To address this, we propose CSSDet, emphasizing multi-scale object detection and small object enhancement. CSSDet integrates an additional detection head for small objects and employs a bidirectional weighted feature pyramid network for effective cross-scale feature fusion, enhancing global perceptual capability. Additionally, we introduce coordinate attention and Involution modules to mitigate information loss. Experiments are conducted on two publicly available datasets, VisDrone and UAVDT. Quantitative results from the experiments indicate that the proposed CSSDet achieves optimal performance across a wide range of evaluation metrics, with particular strengths demonstrated in multi-scale object detection and small-size object detection. Qualitative results showcase its remarkable effectiveness in various aspects of object detection, including small object detection, objects at different scales and categories, multi-perspective scenarios, and complex scenes. The substantial experimental findings confirm that CSSDet attains comprehensive object detection performance on drone-view images. The source code is available at https://github.com/Gui-Cheng/CSSDet.

A Two-Stream Method for Human Action Recognition Using Facial Action Cues.

Human action recognition (HAR) is a critical area in computer vision with wide-ranging applications, including video surveillance, healthcare monitoring, and abnormal behavior detection. Current HAR methods predominantly rely on full-body data, which can limit their effectiveness in real-world scenarios where occlusion is common. In such situations, the face often remains visible, providing valuable cues for action recognition. This paper introduces Face in Action (FIA), a novel two-stream method that leverages facial action cues for robust action recognition under conditions of significant occlusion. FIA consists of an RGB stream and a landmark stream. The RGB stream processes facial image sequences using a fine-spatio-multitemporal (FSM) 3D convolution module, which employs smaller spatial receptive fields to capture detailed local facial movements and larger temporal receptive fields to model broader temporal dynamics. The landmark stream processes facial landmark sequences using a normalized temporal attention (NTA) module within an NTA-GCN block, enhancing the detection of key facial frames and improving overall recognition accuracy. We validate the effectiveness of FIA using the NTU RGB+D and NTU RGB+D 120 datasets, focusing on action categories related to medical conditions. Our experiments demonstrate that FIA significantly outperforms existing methods in scenarios with extensive occlusion, highlighting its potential for practical applications in surveillance and healthcare settings.