The traditional methods of English text classification have two disadvantages. One is that they cannot fully represent the semantic information of the text. The other is that they cannot fully extract and integrate the global and local information of the text. Therefore, we propose a multi-feature fusion model based on long and short term memory network and improved artificial bee colony algorithm for English text classification. In this method, the character-level vector and word-level vector representations of English text are calculated using a pre-training model to obtain a more comprehensive text feature vector representation. Then the multi-head attention mechanism is used to capture the dependencies in the text sequence to improve the semantic understanding of the text. Through feature fusion, the channel features are optimized and the spatial features and time series features are combined to improve the classification performance of the hybrid model. In the stage of network training, the weighted linear combination of maximum Shannon entropy and minimum cross entropy is used as the return degree evaluation function of the bee colony algorithm, and the scale factor is introduced to adjust the solution search strategy of leading bees and following bees, and the improved artificial bee colony algorithm is combined with the classification network to realize the automatic optimization and adjustment of network parameters. Experiments are carried out on public data set. Compared with traditional convolutional neural networks, the classification accuracy of the new model increases by 2% on average, and the accuracy of data set increases by 2.4% at the highest.

Multicriteria clustering problem has been studied and applied scarcely. When a multicriteria clustering problem is tackled with an outranking approach, it is necessary to include preferences of decision makers on the raw dataset, e.g., weights and thresholds of the evaluation criteria. Then, it is necessary to conduct a process to obtain a comprehensive model of preferences represented in a fuzzy or crisp outranking relation. Subsequently, the model can be exploited to derive a multicriteria clustering. This work presents an exhaustive search approach using a credibility similarity measure to exploit a fuzzy outranking relation to derive a multicriteria clustering. The work includes two experimental designs to evaluate the performance of the algorithm. Results show that the proposed method has good performance exploiting fuzzy outranking relations to create the clusterings.

Underwater image denoising technology is of great significance in underwater operation. Underwater operations (such as offshore oil drilling, undersea tunnels, pipeline construction, underwater archaeology, biological research, and lifesaving) require stable and clear underwater images to aid analysis. Due to the scattering and absorption of light by water bodies, obtaining high-quality underwater images is a challenging task. Underwater images are prone to low contrast, low resolution and edge distortion. Therefore, it is difficult to accurately separate the effective signal when removing the underwater image noise, which leads to the image contrast reduction. Also the edge contour is not clear, and the detail loss is serious. Therefore, we propose a novel underwater image denoising method based on curved wave filter and two-dimensional variational mode decomposition. Firstly, the noisy image is decomposed by two-dimensional variational mode decomposition, and a series of modal components with different center frequencies are obtained. The effective modal components are selected by correlation coefficient and structural similarity. And the effective modal components are processed by the curve-wave filter. Finally, the filtered modal components are reconstructed to remove the noise in the image. The experimental results show that, compared with other state-of-the-art methods, the proposed method has clearer denoising results, less mean square error, and better denoising effect.

Animal intelligent breeding utilizes advanced technology and intelligent systems to monitor, analyze, and optimize animal growth environments and management, which can enhance breeding efficiency and animal health levels. In this paper, we propose a new multi-object real-time tracking within deep framework for intelligent breeding of animal (MRT-IB), which consists of semantic feature extraction module, center point prediction module, and object and trajectory calibration module. MRT-IB reduces the difficulty of modeling animal trajectories by performing animal detection on consecutive frames, resulting in higher robustness in real farming scenarios compared to traditional multi-object tracking schemes that directly model animal motion trajectories.

Recent work has demonstrated the Transformer model is effective for computer vision tasks. However, the global self-attention mechanism utilized in Transformer models does not adequately consider the local structure and details of images, which may result in the loss of information and local details, causing decreased estimation accuracy in gaze estimation tasks when compared to convolution or sequential stacking methods. To address this issue, we propose a parallel CNNs-Transformer aggregation network (CTA-Net) for gaze estimation, which fully leverages the advantages of the Transformer model in modeling global context while the convolutional neural networks (CNNs) model in retaining local details. Specifically, Transformer and ResNet are deployed to extract facial and eye information, respectively. Additionally, an attention cross fusion (ACFusion) Block is embedded with CNN branch, which decomposes features in space and channels to supplement lost features, suppress noise, and help extract eye features more effectively. Finally, a dual-feature aggregation (DFA) module is proposed to effectively fuse the output features of both branches with the help feature a selection mechanism and a residual structure. Experimental results on the MPIIGaze and Gaze360 datasets demonstrate that our CTA-Net achieves state-of-the-art results.

Since its emergence at the end of 2019, SARS-CoV-2 has infected millions of people worldwide, putting healthcare systems to the test. This has led many researchers to investigate how machine learning can assist clinicians in diagnosing infections caused by SARS-CoV-2. Building upon prior studies, we propose a novel deep learning framework designed for the segmentation of lesions evident in Computed Tomography (CT) scans. For this work, we utilized a dataset consisting of 20 CT scans annotated by experts and conducted training, validation and external evaluation of the deep learning models we implemented, employing a 5-fold cross-validation scheme. When splitting data by slice, our optimal model achieved noteworthy performance metrics, attaining a Dice Similarity Coefficient (DSC) and Intersection over Union (IoU) score of 0.8644 and 0.7612, respectively, during the validation phase. In the external evaluation phase, the model maintained a strong performance with a DSC and IoU score of 0.7211 and 0.5641, respectively. When splitting data by patient, our optimal model achieved a DSC and IoU score of 0.7889 and 0.6540, respectively, during the validation phase. In the external evaluation phase, the model maintained a strong performance with a DSC and IoU score of 0.7079 and 0.5480, respectively. The results, obtained from this research, suggest that the incorporation of transfer learning, combined with appropriate pre-processing techniques, can contribute to achieving state-of-the-art performance in the segmentation of lesions associated with SARS-CoV-2 infections.

The advent of newer and better technologies has made Human Activity Recognition (HAR) highly essential in our daily lives. HAR is a classification problem where the activity of humans is classified by analyzing the data collected from various sources like sensors, cameras etc. for a period of time. In this work, we have proposed a model for activity recognition which will provide a substructure for the assisted living environment. We used a genetic search based feature selection for the management of the voluminous data generated from various embedded sensors such as accelerometer, gyroscope, etc. We evaluated the proposed model on a sensor-based dataset - Human Activities and Postural Transitions Recognition (HAPT) which is publically available. The proposed model yields an accuracy of 97.04% and is better as compared to the other existing classification algorithms on the basis of several considered evaluation metrics. In this paper, we have also presented a cloud based edge computing architecture for the deployment of the proposed model which will ensure faster and uninterrupted assisted living environment.

The consumption of energy and carbon emission in cloud datacenters are the alarming issues in recent times, while optimizing the average response time and service level agreement (SLA) violations. Handful of researches have been conducted in these domains during virtual machine placement (VMP) in cloud milieu. Moreover it is hard to find researches on VMP considering the cloud regions and the availability zones along with the datacenters, although both of them play significant roles in VMP. Hence, we have worked on a novel approach to propose a hybrid metaheuristic technique combining the salp swarm optimization and emperor penguins colony algorithm, i.e. SSEPC to place the virtual machines in the most suitable regions, availability zones, datacenters, and servers in a cloud environment, while optimizing the mentioned quality of service parameters. Our suggested technique is compared with some of the contemporary hybrid algorithms in this direction like Sine Cosine Algorithm and Salp Swarm Algorithm (SCA-SSA), Genetic Algorithm and Tabu-search Algorithm (GATA), and Order Exchange & Migration algorithm and Ant Colony System algorithm (OEMACS) to test its efficacy. It is found that the proposed SSEPC is consuming 4.4%, 8.2%, and 16.6% less energy and emitting 28.8%, 32.83%, and 37.45% less carbon, whereas reducing the average response time by 11.43%, 18.57%, and 26% as compared to its counterparts GATA, OEMACS, and SCA-SSA respectively. In case of SLA violations, SSEPC has shown its effectiveness by lessening the value of this parameter by 0.4%, 1.2%, and 2.8% as compared to SCA-SSA, GATA, and OEMACS respectively.

Pedestrian attribute recognition is an important task for intelligent video surveillance. However, existing methods struggle to accurately localize discriminative regions for each attribute. We propose Attribute Localization Transformer (ALFormer), a novel framework to improve spatial localization through two key components. First, we introduce Mask Contrast Learning (MCL) to suppress regional feature relevance, forcing the model to focus on intrinsic spatial areas for each attribute. Second, we design an Attribute Spatial Memory (ASM) module to generate reliable attention maps that capture inherent locations for each attribute. Extensive experiments on two benchmark datasets demonstrate state-of-the-art performance of ALFormer. Ablation studies and visualizations verify the effectiveness of the proposed modules in improving attribute localization. Our work provides a simple yet effective approach to exploit spatial consistency for enhanced pedestrian attribute recognition.

Currently, there is relatively little formal analysis and verification work on the 5G EAP-TLS authentication protocol. In this paper, we use the model checker SPIN to perform a formal analysis of the 5G EAP-TLS authentication protocol. Firstly, we analyze the process of the 5G EAP-TLS authentication protocol and abstract it to obtain a formal model of the protocol. Then, we describe the construction of the protocol model based on the Promela language. The unique feature of this paper is the replacement of the hash value of the 5G EAP-TLS authentication protocol with the message content field encrypted by an unknown subject public key. This is because the Promela language in SPIN has an eval function that can check the value of each field. This can replace the function of the hash function and make the Promela model construction more portable. The paper analyzes the attack paths of the protocol and reveals design flaws that undermine the expected identity authentication attributes and secret consistency of the protocol. The results not only provide a comprehensive understanding of the security properties of the 5G EAP-TLS authentication protocol but also offer valuable insights and guidance for the verification of the protocol's security properties, security design, and optimization of protocol implementation and interoperability.