Abstract Identifying and preventing malicious network behavior is a challenge for establishing secure communication environment or system. Malicious activities in system can seriously threaten users’ privacy potentially jeopardize the entire infrastructure functions. Furthermore, cyber-attacks have grown complexity number due to ever-evolving digital landscape of computer devices recent years. Analyzing traffic using intrusion detection systems (NIDSs) has become an integral security measure...

The COVID-19 pandemic has had a significant impact on many lives and the economies of countries since late December 2019. Early detection with high accuracy is essential to help break chain transmission. Several radiological methodologies, such as CT scan chest X-ray, have been employed in diagnosing monitoring disease. Still, these methodologies are time-consuming require trial error. Machine learning techniques currently being applied by several studies deal COVID-19. This study exploits...

Natural disasters, such as floods, can cause significant damage to both the environment and human life. Rapid accurate identification of affected areas is crucial for effective disaster response recovery efforts. In this paper, we aimed evaluate performance state-of-the-art (SOTA) computer vision models flood image classification, by utilizing a semi-supervised learning approach on dataset named FloodNet. To achieve this, trained son 11 modified them suit classification task at hand....

In hyperspectral image (HSI) classification, Convolutional Neural Networks (CNNs) have exhibited exceptional performance, owing to their hierarchical nonlinear modeling. However, fixed square receptive field constrains ability effectively handle irregular regions. Graph Convolution (GCNs) been introduced learn regions through correlations between adjacent pixels modeled as superpixel-based nodes, yet they lack pixel-level information. We propose a novel approach "Pixel-level with Covariance...

With the procession of technology, human-machine interaction research field is in growing need robust automatic emotion recognition systems. Building machines that interact with humans by comprehending emotions paves way for developing systems equipped human-like intelligence. Previous architecture this often considers RNN models. However, these models are unable to learn in-depth contextual features intuitively. This paper proposes a transformer-based model utilizes speech data instituted...

The unparalleled availability of Satellite Image Time Series (SITS) for crop phenology classification unravels agricultural parcel observation and monitoring with applications both economic ecological importance. Moreover, the need distinct parcels into individual types falls on state-of-the-art deep learning models this extrinsic task. However, most existing approaches implemented are complex ineffective attention incorporated models, which in turn lack resilience to recognize useful bands...

The COVID-19 virus has rapidly spread as a global pandemic, causing significant impacts on public health, economies, and daily life worldwide. Accurately quickly predicting is crucial to maintaining stronger healthcare systems. This paper introduces novel hybrid model of artificial intelligence that combines the benefits Variational Auto-Encoder (VAE) with attention mechanism based Vision Transformer (Vi <tex xmlns:mml="http://www.w3.org/1998/Math/MathML"...

Emotion recognition is a topic of significant interest in assistive robotics due to the need equip robots with ability comprehend human behavior, facilitating their effective interaction our society. Consequently, efficient and dependable emotion systems supporting optimal human-machine communication are required. Multi-modality (including speech, audio, text, images, videos) typically exploited tasks. Much relevant research based on merging multiple data modalities training deep learning...

Urinary tract infections (UTIs) are frequent hospital-acquired infection, with Escherichia coli and Proteus mirabilis accounting for 90% of complicated UTIs. Emergence multidrug-resistant (MDR) bacteria have the treatment P. related UTIs has been associated production urinary stones long-term in patients catheters. other uropathogens constitute a largely unexplored pathogen group. The is resistant to most antibiotics as result its impermeable outer membrane (OM). β-barrel assemble machinery...

Emotion recognition is a topic of significant interest in assistive robotics due to the need equip robots with ability comprehend human behavior, facilitating their effective interaction our society. Consequently, efficient and dependable emotion systems supporting optimal human-machine communication are required. Multi-modality (including speech, audio, text, images, videos) typically exploited tasks. Much relevant research based on merging multiple data modalities training deep learning...

Machine learning methods based on fully convolutional networks have emerged as a viable choice for retinal vessel segmentation. However, when input samples significantly deviate from the training data distribution, these deep fail silently, posing serious challenges automatic processing pipelines. In our study, we question effectiveness of randomly selecting unlabeled target images annotation and inclusion in support set, it may not facilitate an efficient fine-tuning process. We propose...

There is an urgent demand for lightweight deep learning models applicable to real-world scenarios. This research proposes innovative XAI model that integrates attention mechanism with multiscale kernel depth-wise separable convolution designed COVID-19 classification using chest X-rays. The consists of four sequential blocks, incorporating (MKnADSC) modules. Notably, this achieves impressive accuracy 96.50%, boasting a compact structure 2.56 million parameters and FLOPs 0.41 G. These...

Training convolutional neural networks (CNNs) for semi-semantic segmentation in remote sensing images is difficult due to the substantial amount of labelled data required. The scarcity dense pixel-level annotations hampers model's ability extract meaningful information from images. We propose a novel semi-supervised semantic network called 3S-Net, designed explicitly high-resolution imagery overcome this challenge. This leverages limited number examples and many unlabeled generate...

The paper mainly studies methods on lane line detection in autonomous driving applications. specific aim of is to experiment the instance-based segmentation technique for a images. We incorporate process selecting portions or patches pixels as keypoints. then improve proposed method with deep metric learning techniques, more specifically, an associative embedding having angular loss function, ensure that, keypoints similar features belongs same instance line. Next, completed by cluster...

Malware attacks in the cyber world continue to increase despite efforts of analysts combat this problem. Recently, samples have been presented as binary sequences and assembly codes. However, most researchers focus only on raw sequence their proposed solutions, ignoring that codes may contain important details enable rapid detection. In work, we leveraged capabilities deep autoencoders investigate presence feature disparities samples. First, treated task outliers whether autoencoder would...