Computer Science
Lightweight Research on Fatigue Driving Face Detection Based on YOLOv8
With the rapid development of society motor vehicles have become one of the main means of transportation. However as the number of motor vehicles continues to increase traffic safety accidents also continue to appear bringing serious threats to people's lives and property safety. Fatigue driving is one of the important causes of traffic safety accidents.
To address this problem a target detection algorithm called VA-YOLO is designed to improve the speed and accuracy of facial recognition for fatigue checking. The algorithm employs a lightweight backbone network VanillaNet instead of the traditional backbone network which reduces the computational and parametric quantities of the model. The SE attention mechanism is also introduced to enhance the model's attention to the target features which further improves the accuracy of target detection. Finally in terms of the bounding box regression loss function the SIoU loss function is used to reduce the error.
The experimental results show that compared to YOLOv8n the VA-YOLO algorithm improves the accuracy by 1.3% while the number of parameters decreases by 30%.
This shows that the VA-YOLO algorithm has a significant advantage in realizing the balance between the number of parameters and accuracy which is important for improving the speed and accuracy of fatigue driving detection.
Patent Selections
Revolutionizing Industries Through AI-Driven Robotics
A Brief Review on Recent Advancements of MEMS Devices for Telecommunication Applications
This review offers a detailed analysis of recent advancements in Micro-Electro-Mechanical Systems (MEMS) technology and its impact on telecommunication applications. MEMS devices have undergone significant evolution with enhancements in performance and functionality crucial for modern telecommunication systems. The review covers various advancements including developing MEMS-based switches filters and oscillators—key developments with sophisticated features including the equipment to enable modern cost-effective communication networks. Infrastructure equipment such as passive elements tunable networks antennas etc. with radio frequency (RF) as well as optical devices and mobile communication devices i.e. the mobile sensors and actuators are the two main areas encompassing this field of application because of improved performance and quality of user experience. These components are essential for managing high-frequency signals improving signal quality and supporting the demands of increasingly complex telecommunication networks. Recent innovations have enabled MEMS devices to operate with higher precision and reliability making them integral to the rollout of 5G networks and other emerging technologies. The review also addresses the challenges faced in the field such as issues related to device miniaturization cost-effectiveness and the need for seamless integration with existing infrastructure. By examining these advancements and challenges this review provides a comprehensive overview of the current state of MEMS technology in telecommunications. It aims to serve as a valuable resource for researchers engineers and industry professionals seeking to understand MEMS devices' recent progress and future potential in enhancing telecommunication systems. The insights provided herein are intended to guide ongoing research and development efforts ultimately contributing to the continued advancement of telecommunication technologies.
Unique Taxonomy and Review of New Age Smart Home IoT Forensics Tools
In the field of digital forensics the proliferation of Internet of Things (IoT) devices within intelligent residences has presented both new opportunities and challenges. Every gadget including smart thermostats security cameras lighting controls and even washing machines or refrigerators is equipped by these installers in the range of gadgets offered by manufacturers.
This research conducts a comprehensive investigation and classification of contemporary forensic tools designed for IoT devices in smart homes highlighting their characteristics methods of data collection types of target devices analytical methodologies practical applications and capabilities for integration. This generally involves a comparison of different forensic products or solutions and evaluation on various criteria such as cost supportability maintainable architecture-designs (integration) speedy acquisition speed/performance effectiveness without compromising quality ease-of-use and consistency.
A comparative analysis with detailed tables & radar charts identifying the detailed pros and cons of each tool our findings help forensic professionals understand when to use them for effective decisions. The results show that XRY and UFED by Cellebrite both scored 5/5 in each criterion showing the best performance in mobile device forensics. Wireshark and tcpdump also have high rates for the accuracy and reliability criteria with results of 5/5 and are therefore also highly recommended in the area of analysis of network traffic. Magnet AXIOM and NetworkMiner graded evenly well with a usability rating of four out of five and an integration mark of 4 out of five which diversified them for computer and mobile forensics. Splunk and ELK Stack scored topping the scalability category with each scoring out of five which further confirmed the analysis of logs well for large data sets. These numerical results further underline that the choice of the tool depends on specific forensic requirements.
The authors examine future IoT forensics in smart homes which highlights the necessity of devices working with each other through a standard and sophisticated analysis to deal with dynamic complexity development within this field.
Comprehensive Analysis of Oversampling Techniques for Addressing Class Imbalance Employing Machine Learning Models
Unbalanced datasets present a significant challenge in machine learning often leading to biased models that favor the majority class. Recent oversampling techniques like SMOTE Borderline SMOTE and ADASYN attempt to mitigate these issues. This study investigates these techniques in conjunction with machine learning models like SVM decision tree and logistic regression. The results reveal critical challenges such as noise amplification and overfitting which we address by refining the oversampling approaches to improve model performance and generalization.
In order to address this challenge of unbalanced datasets the minority class is oversampled to accommodate the majority class. Oversampling techniques such SMOTE (Synthetic Minority Oversampling Technique) Borderline SMOTE and ADASYN (Adaptive Synthetic Sampling) are used in this work.
To perform the comprehensive analysis of various oversampling methods for taking care of class imbalance issue using ML methods.
The proposed methodology uses BERT technique which removes the pre-processing step. Various proposed oversampling techniques in the literature are used for balancing the data followed by feature extraction and text classification using ML algorithms. Experiments are performed using ML classification algorithms like Decision tree (DT) Logistic regression (LR) Support vector machine (SVM) and Random forest (RF) for categorizing the data.
The results show improvement corresponding SVM using Borderline SMOTE resulting in an accuracy of 71.9% and MCC value of 0.53.
The suggested method assists in the evolution of fairer and more effective ML models by addressing this basic issue of class imbalance.
IoT-Aided Self-Adaptive Routing for Flying Ad Hoc Network Using Gauss-Markov Mobility Model
Unmanned aerial vehicles (UAVs) are low-cost easy to deploy and can be integrated with IoT to solve human-related problems. They have applications in military and civil sectors such as forest fire detection sports monitoring agriculture and border surveillance. UAV-based applications can be signal or multi-system.
This work is an attempt to present routing protocols within FANET. AntHocNet has shown better results than traditional routing protocols. While basic principle of ant colony optimization is used in the proposed algorithm. Engineers face a lot of problems due to the dynamic behavior of UAVs. Traditional routing protocols are compared with the proposed algorithm in simulation results. Gauss-Markov mobility model is used which easily covers temporal dependencies. Quality of experience parameters are utilized to check routing protocols' performance. More interestingly various applications of UAVs are discussed in detail.
Unbalanced communication in UAVs directly disturbs the entire network.
Therefore routing protocols are introduced in UAV-to-UAV communication. During simulation packet loss throughput end-to-end delay bandwidth utilization packet drop rate and packet delivery parameters are used.
Field Pest Detection via Pyramid Vision Transformer and Prime Sample Attention
Pest detection plays a crucial role in smart agriculture; it is one of the primary factors that significantly impact crop yield and quality.
In actual field environments pests often appear as dense and small objects which pose a great challenge to field pest detection. Therefore this paper addresses the problem of dense small pest detection.
We combine a pyramid vision transformer and prime sample attention (named PVT-PSA) to design an effective pest detection model. Firstly a pyramid vision transformer is adopted to extract pest feature information. Pyramid vision transformer fuses multi-scale pest features through pyramid structure and can capture context information of small pests which is conducive to the feature expression of small pests. Then we design prime sample attention to guide the selection of pest samples in the model training process to alleviate the occlusion effect between dense pests and enhance the overall pest detection accuracy.
The effectiveness of each module is verified by the ablation experiment. According to the comparison experiment the detection and inference performance of the PVT-PSA is better than the other eleven detectors in field pest detection. Finally we deploy the PVT- PSA model on a terrestrial robot based on the Jetson TX2 motherboard for field pest detection.
The pyramid vision transformer is utilized to extract relevant features of pests. Additionally prime sample attention is employed to identify key samples that aid in effectively training the pest detection models. The model deployment further demonstrates the practicality and effectiveness of our proposed approach in smart agriculture applications.
Discovery of Two GSK3β Inhibitors from Sophora flavescens Ait. using Structure-based Virtual Screening and Bioactivity Evaluation
Kushen (Sophora flavescens Ait.) has a long history of medicinal use in China due to its medicinal values such as antibacterial antiviral and anti-inflammatory. Rapid discovery of the components and the medicinal effects exerted by Kushen will help elucidate the science of Kushen in curing diseases. GSK3β (glycogen synthase kinase-3 beta) is a protein kinase with a wide range of physiological functions such as antibacterial antiviral and anti-inflammatory. The discovery of inhibitors targeting GSK3β from Kushen was not only helpful for the rapid discovery of the components responsible for the efficacy of Kushen but also important for the development of novel drugs.
In this study the chemical composition of Kushen was extracted from the TMSCP database. Molecular docking GSK3β enzyme assay and molecular dynamics simulations were used to discover the GSK3β inhibitors from the chemical composition of Kushen.
A total of 113 chemical compositions of Kushen were extracted from the TMSCP database. Molecular docking indicated that 15 chemical compositions of Kushen scored better than -8 kcal/mol against GSK3β. GSK3β enzyme assay demonstrated several inhibitory activities of kushenol I and kushenol F with IC50 values of 7.53 ± 2.55 µM and 4.96 ± 1.29 µM respectively. Molecular dynamics simulations were used to reveal the interactions of kushenol I and kushenol F with GSK3β from structural and energetic perspectives.
Kushenol I and kushenol F could be the material basis for the antibacterial antiviral and anti-inflammatory properties of Kushen.
Exploring the Mechanism of Centipeda minima in Treating Nasopharyngeal Carcinoma Based on Network Pharmacology
Centipeda minima (CM) is a traditional Chinese herbal medicine used for the treatment of sinusitis and rhinitis and it possesses anti-cancer properties. However the mechanism of CM in the treatment of nasopharyngeal carcinoma (NPC) remains unclear.
This study aimed to explore the mechanism of CM in the treatment of NPC using a network pharmacology approach.
The active components and targets of CM and NPC were screened using TCMSP SwissTarget and GeneCards database. The association between CM components and NPC targets or pathways was analyzed using String Cytoscape 3.9.1 David 6.7 and AutoDock Vina. The Sangerbox platform was used to conduct differential expression and Kaplan-Meier survival analysis of core genes.
We identified 17 active compounds of CM and 146 corresponding targeted proteins in NPC. These targets may modulate pathways in cancer PI3K-Akt apoptosis prolactin relaxin and TNF signaling. The top 5 core genes of the PPI network were found to be AKT1 STAT3 CASP3 EGFR and SRC which may be the main targets of CM in treating NPC. Molecular docking confirmed the binding energies of quercetin with CASP3 8-Hydroxy-910-diisobutyryloxythymol with AKT1 and plenolin with AKT1 which were particularly low suggesting robust and stable interactions. The expression levels of AKT1 CASP3 EGFR SRC MMP9 PTGS2 are significantly higher in head and neck squamous cell carcinoma (HNSC) samples compared to normal samples. In addition the hub genes could predict the prognosis of HNSC as the Kaplan-Meier survival curve showed that patients with lower expressions of AKT1 EGFR SRC CCND1 PPARG had better overall survival.
By conducting a network pharmacology approach we revealed the main ingredients key targets and regulatory pathways of Centipeda minima in the treatment of NPC.
Graph Convolutional Network and Attention for Traffic Prediction - A Deep Learning Approach
Traffic prediction is a key component of the intelligent transportation system for researchers and practitioners. It is extremely challenging because traffic flows typically exhibit complicated patterns complex spatio-temporal correlations and non-linearities.
Prediction of traffic density on the roads can help not only urban traffic management but also support for other road services such as path planning.
In this paper we propose an attention-based graph convolutional network (AGCN) model to solve the traffic prediction problem. The primary focus of AGCN is on temporal daily and weekly dependencies of traffic periodicity. To efficiently capture the dynamic geographical and temporal correlations in traffic data an attention-based spatial-temporal mechanism is employed. Additionally standard convolutions are employed to extract temporal data and graph convolutional networks are used to capture spatial patterns.
The final prediction results are generated by fusing the outputs of these components. California Transportation Agencies Performance Measurement System (CalTrans PeMS) dataset is used in this research to assess the performance. The proposed model has been validated using simulations that exhibit the viability of the method and show 4% increase in the accuracy of prediction.
For improved route planning and to arrive at the destination in the least amount of time an efficient traffic pre- diction model is suggested. This enhances overall transportation system efficiency and aids in traffic control.
Discovery of Novel PTP1B Inhibitors by High-throughput Virtual Screening
To Discover novel PTP1B inhibitors by high-throughput virtual screening.
Type 2 Diabetes is a significant global health concern. According to projections the estimated number of individuals affected by the condition will reach 578 million by the year 2030 and is expected to further increase to 700 million deaths by 2045. Protein Tyrosine Phosphatase 1B is an enzymatic protein that has a negative regulatory effect on the pathways involved in insulin signaling. This regulatory action ultimately results in the development of insulin resistance and the subsequent elevation of glucose levels in the bloodstream. The proper functioning of insulin signaling is essential for maintaining glucose homeostasis whereas the disruption of insulin signaling can result in the development of type 2 diabetes. Consequently we sought to utilize PTP1B as a drug target in this investigation.
The purpose of our study was to identify novel PTP1B inhibitors as a potential treatment for managing type 2 diabetes.
To discover potent PTP1B inhibitors we have screened the Maybridge HitDiscover database by SBVS. Top hits have been passed based on various drug-likeness rules toxicity predictions ADME assessment Consensus Molecular docking DFT and 300 ns MD Simulations.
Compound RJC02059 has been identified with strong binding affinity at the active site of PTP1B along with drug-like properties efficient ADME low toxicity and high stability.
Two compounds demonstrated strong binding affinity favorable drug-like properties and stable interactions with PTP1B's active site throughout 200 ns MD simulations with RJC02059 showing superior binding stability and persistent hydrogen bonding with catalytic residues. However experimental validation through enzymatic assays and assessment of selectivity against related phosphatases remain essential next steps to confirm therapeutic potential.
The identified molecule could potentially manage T2DM effectively by inhibiting PTP1B providing a promising avenue for therapeutic strategies.
Identifying Novel Inhibitors for Dengue NS2B-NS3 Protease by Combining Topological similarity, Molecular Dynamics, MMGBSA and SiteMap Analysis
DENV NS2B-NS3 protease inhibitors were designed based upon the reference molecule 4-(13-dioxoisoindolin-2-yl)-N-(4-ethylphenyl) benzenesulfonamide reported by our team with the aim to optimize lead compound via rational approach. Top five best scoring molecules with zinc ids ZINC23504872 ZINC48412318 ZINC00413269 ZINC13998032 and ZINC75249613 bearing ‘pyrimidin-4(3H)-one’ basic scaffold have been identified as a promising candidate against DENV protease enzyme.
The shape and electrostatic complementary between identified HITs and reference molecules were found to be Tanimotoshape 0.453 0.690 0.680 0.685 & 0.672 respectively and Tanimotoelectrostatic 0.211 0.211 0.441 0.442 0.442 and 0.442 respectively. The molecular docking studies suggested that the identified HITs displayed the good interactions with active site residues and lower binding energies. The stability of docked complexes was assessed by MD simulations studies. The RMSD values of protein backbone (1.6779 3.1563 3.3634 3.3893 & 3.0960 Å) and protein backbone RMSF values (1.0126 1.0834 1.0890 0.9974 & 1.0080 Å respectively) for all top five HITs were stable and molecules did not fluctuate from the active pocket during entire 100ns MD run.
The druggability Dscore below 1 indicate the tightly binding of ligand at the active site. Dscore for ZINC23504872 was found to be 1.084 while for the second class of compounds ZINC48412318 ZINC00413269 ZINC13998032 and ZINC75249613 0.503 0.484 0.487 and 0.501 Dscores were observed. In-silico ADMET calculations suggested that all five HITs were possessed the drug likeliness properties and did not violate the Lipinski’s rule of five.
Summing up these in-silico generated data suggested that the identified molecules bearing pyrimidin-4(3H)-one would be promising scaffold for DENV protease inhibitors. However experimental results are needed to prove the obtained results.
Exploring the Potential Mechanisms of Danshen for the Treatment of Ulcerative Colitis based on Serum Pharmacochemistry, Gene Expression Profiling, and Network Pharmacology: Regulation of Cell Apoptosis and Inflammatory Response
As a traditional Chinese medicine Danshen shows potential efficacy for treating ulcerative colitis (UC). However the bioactive components and mode of action were unclear.
This paper uses a combination of network pharmacology serum medicinal chemistry and gene expression profiling to clarify its possible molecular mechanism of action and material basis.
Ultra-high performance liquid chromatography-mass spectrometry (UPLC-MS) was utilized to analyze the herbal components and metabolites from the serum of Danshen-treated mice. Gene expression profiles were applied to construct a database of Danshen action targets. Then active ingredient-target-biological functional module networks were constructed to analyze the mechanism of action. Molecular docking has further confirmed the possibility of its components to the targets.
As a result 193 common targets between 1684 Danshen-related DEGs and 1492 UC targets were determined as the potential targets for Danshen in treatment with UC. Serum pharmacochemistry and target prediction showed that 22 components in serum acted on 777 targets. Intersection with common targets yielded 46 core targets and an active ingredient-target-biological functional module network was constructed for analysis. Network prediction and molecular docking results showed that the main action modules were inflammatory response and cell apoptosis which mainly acted on targets SRC RELA HSP90AA1 CTNNB1 STAT3 and CASP3. The main components of Danshen intervention in UC were predicted to include Catechol 39-Dimethoxypterocarpan 8-Prenylnaringenin Isoferulic acid Salvianolic acid C and Danshensu.
This work resolved the ambiguity of Danshen’s anti-UC material basis and mechanism: 22 serum-absorbed components acted on 46 core targets to modulate inflammation and apoptosis validating the integrated approach and laying groundwork for UC treatment and TCM research.
The present study provides a scientific foundation for further explicating the mechanisms of Danshen against UC.