A. Usha Ruby

Background and purpose: Fetal ultrasound image segmentation is a crucial component of prenatal diagnosis, enabling the accurate measurement of key biometric parameters and the early detection of abnormalities. Manual segmentation is prone to variability due to factors such as operator skill, image quality, and acquisition conditions. This study compares conventional and deep learning-based segmentation methods, proposing an enhanced Multi-Feature Pyramid U-Net (MFP-Unet) to improve accuracy, robustness, and clinical applicability. Methods: We utilised the HC18 Grand Challenge Dataset, which comprises 1,334 ultrasound images from 551 pregnancies without growth abnormalities. The proposed MFP-Unet integrates multi-scale feature pyramid fusion, dilated convolutions, and attention gates into the U-Net framework. Performance was compared with Dilated U-Net, U-Net, RU-Net, Attention U-Net, and R2U-Net using Dice Coefficient, Intersection over Union (IoU), and mean Average Precision (mAP). Preprocessing included Niblack’s adaptive thresholding and custom region extraction for femur length images. Results: MFP-Unet achieved the highest accuracy across all parameters: HC IoU = 90% (Dice = 0.95), FL IoU = 93% (Dice = 0.97), BPD IoU = 88% (Dice = 0.93), and overall mAP = 0.92. The architecture formed smoother anatomical boundaries, enhanced robustness to image variability, and compact measurement variations associated with other methods. Conclusions: The MFP-Unet approach shows strong promise for use in clinical fetal ultrasound analysis, offering reliable and precise biometric measurements that can aid in timely interventions and enhance prenatal care outcomes. Future research should validate performance on multi-centre datasets and explore real-time deployment.

Industrial Wireless Sensor Networks (IWSNs) play a vital role in today’s Industrial Internet of Things (IoT) systems. The existing Enhanced Energy Optimization Model has made progress in improving energy use; however, it still faces issues with adapting to changing network traffic, slow clustering convergence, and limited optimization capabilities. To address these problems, this paper presents a hybrid intelligent framework that improves energy savings, clustering convergence, and adaptive network optimization in IWSNs. The new model combines Enhanced Aphid–Ant Mutualism Optimization (EAAM) for efficient clustering, an Improved Artificial Bee Colony (I-ABC) algorithm for ongoing energy parameter adjustments, and a Graph Neural Network (GNN)-based predictor for adapting to changing topologies and traffic conditions. EAAM provides fair cluster-head selection and minimizes control overhead, while I-ABC adjusts transmission power and duty cycles to reduce energy loss. The GNN module captures spatial correlations between sensor nodes to provide proactive energy prediction and adaptive reconfiguration, with lower training overhead than deep reinforcement learning. Experimental results on benchmark IWSN datasets verify that the proposed EAAM–I-ABC–GNN framework exhibits 18%–25% better overall energy efficiency, a 22% increase in network lifetime, and 15% improved clustering stability compared to the baseline EEOM and traditional hybrid methods. These findings validate that the proposed hybrid model substantially improves scalability, flexibility, and sustainability, providing an efficient and intelligent solution for future Industrial IoT networks.

American scientific publishing group

Face recognition is a method of biometric identification technology that uses a person's face characteristic data. Face-based characteristics can be easily acquired, unlike fingerprints, iris, and other biometrics, they can be collected without physical contact. Face recognition technology has therefore rapidly grown and is now widely employed in a variety of industries, including banking, manufacturing, banking, healthcare, and education. Convolutional neural networks (CNNs) have attained good results in face recognition with the constant developments in deep learning. However, throughout the training phase, deep convolution neural networks have challenges in convergence and optimization. These issues are resolved by residual networks. Furthermore, the channel attention techniques can support networks in learning only the characteristics that carry valuable information, hence enhancing the model’s accuracy. In this study, we first use the swish function to enhance the initial residual network to produce the improved residual network and then the spatial and channel attention mechanisms, are added to the Resnet. The experimental findings of face recognition on MegaFace, CFP, LFW, and AgeDB datasets demonstrate that our model performed significantly in various metrics like specificity, sensitivity, test accuracy, F1 score, Matthew's correlation coefficient, precision, and test loss.

Phishing attacks remain an enduring and ever-evolving menace to both networked systems and their users' privacy. In response to this formidable challenge, our research delves into an innovative approach designed to enhance the precision of phishing Uniform Resource Locator (URL) detection within the dynamic and programmable realm of Software-Defined Networks (SDNs). By harnessing feature selection capabilities and adaptive machine learning techniques, our proposed framework aims to fortify security measures in SDNs against these malicious campaigns. Our methodology's core is the deliberate selection of discriminative features from the extensive network data attributes. This feature selection process is meticulously designed to identify the most relevant characteristics associated with phishing URLs, thereby enabling the extraction of invaluable insights for more precise detection. These carefully chosen features then serve as inputs for a dynamic machine learning model, trained to adapt and evolve alongside the constantly changing landscape of phishing attacks. Within the SDN environment, our framework optimizes utilizing network resources and controller processing power. It achieves this by reducing the dimensionality of input data, resulting in improved detection accuracy and a decrease in false positives. The adaptive nature of our machine learning model ensures rapid recognition of emerging phishing tactics, thereby reducing the risk of succumbing to novel and sophisticated attacks. To validate the effectiveness of our approach, we conducted extensive experiments and evaluations within an SDN testbed, utilizing real-world phishing URL datasets. The results consistently demonstrate that our framework surpasses conventional methods, achieving higher detection accuracy and adaptability to evolving threats. In summary, our research represents a significant stride in the ongoing battle against phishing attacks by leveraging the dynamic capabilities of SDNs. The synergy between feature selection and adaptive machine learning techniques empowers SDNs to sustain accurate and effective phishing URL detection, ultimately reinforcing network security and safeguarding user privacy in an ever-evolving threat landscape.

Spam comments pose a significant challenge in maintaining the quality of online platforms, including YouTube. In this paper, we propose a novel approach, the Graph-Enhanced Hierarchical Attention Network (GE-HAN), for spam comment detection. Our approach leverages the content of comments, the hierarchical structure of conversations, and user interactions to accurately classify comments as spam or non-spam. The GE-HAN model combines attention mechanisms to capture important textual features within comments and graph convolutional networks to analyze user interactions and detect coordinated spamming efforts. We trained and evaluated the model using a labeled dataset of spam comments, incorporating user names along with comment content and adjacency matrix representing user interactions. Experimental results demonstrate that the GE-HAN model achieves superior performance in spam detection, outperforming traditional methods. By considering both content and user dynamics, the model effectively identifies spam comments in YouTube, providing a robust solution to combat spam. Our research contributes to the field by showcasing the potential of graph-enhanced attention mechanisms in capturing complex patterns within comments and user interactions.

This paper compares the performance of various popular convolutional neural network (CNN) architectures for image classification on the CIFAR10 dataset. The comparison includes CNN architectures such as Inception V3, Inception-ResNet-v2, ResNetV1, and V2, ResNeXt, MobileNet, and DenseNet, with the addition of two attention mechanisms - Convolutional Block Attention Module (CBAM), and Squeeze and Excitation (SE). CBAM and SE are believed to improve CNNs' performance, especially for complex images with multiple objects and backgrounds. The models are evaluated using loss and accuracy. The main focus of this study is to identify the most effective CNN architecture for image classification on the CIFAR10 dataset with attention mechanisms. The study aims to compare the accuracy of various CNN architectures with and without attention mechanisms and to identify the critical differences between these architectures in terms of their ability to handle complex images. The findings of this study could have implications for developing advanced CNN architectures that can potentially improve the accuracy of computer vision systems in various applications.