Search Results (300)

Electricity theft remains a pressing challenge in modern smart grid systems, leading to significant economic losses and compromised grid stability. This paper presents a sensor-driven framework for electricity theft detection that leverages data collected from smart meter sensors, key components in smart grid monitoring infrastructure. The proposed approach combines prototype learning and meta-level ensemble learning to develop a scalable and accurate detection model, capable of identifying zero-day attacks that are not present in the training data. Smart meter data is compressed using Principal Component Analysis (PCA) and K-means clustering to extract representative consumption patterns, i.e., prototypes, achieving a 92% reduction in dataset size while preserving critical anomaly-relevant features. These prototypes are then used to train base-level one-class classifiers, specifically the One-Class Support Vector Machine (OCSVM) and the Gaussian Mixture Model (GMM). The outputs of these classifiers are normalized and fused in a meta-OCSVM layer, which learns decision boundaries in the transformed score space. Experimental results using the Irish CER Smart Metering Project (SMP) dataset show that the proposed sensor-based detection framework achieves superior performance, with an accuracy of 88.45% and a false alarm rate of just 13.85%, while reducing training time by over 75%. By efficiently processing high-frequency smart meter sensor data, this model contributes to developing real-time and energy-efficient anomaly detection systems in smart grid environments. Full article

The overuse of synthetic nitrogen fertilizer in vineyards degrades soil quality and poses environmental risks. Partial substitution of synthetic nitrogen with organic alternatives may enhance grapevine performance and soil sustainability, depending on the substitution rate. This study evaluated the effects of replacing synthetic nitrogen with composted spent mushroom substrate at five different rates (0%, 25%, 50%, 75%, and 100%, denoted as NOS, OS-25, OS-50, OS-75, and OS-100, respectively) and a control with no nitrogen fertilization applied (CK), on soil fertility, root growth, and photosynthetic performance in grapevine seedlings. Compared to CK, nitrogen fertilization and organic substitution significantly increased soil electrical conductivity, organic matter, and macronutrient contents, but had no significant effect on soil pH. Organic substitution markedly improved leaf photosynthetic capacity in the summer, with the highest rates observed under OS-25, exceeding CK and NOS by 32.98–63.19% and 13.93–27.38%, respectively. Root growth was also significantly enhanced by organic substitution, with OS-25 exhibiting the best performance. Fine roots in the 0.0–0.5 mm diameter class were dominant, accounting for 56.88–63.06% of total root length and 96.22–97.31% of total root tip count. Increasing substitution rates beyond 25% yielded no further improvements in photosynthesis or root growth. Mantel test analysis indicated strong positive correlations between soil fertility parameters (e.g., alkali-hydrolyzable nitrogen, available phosphorous and potassium) and both photosynthetic efficiency and root growth. These findings suggest that an appropriate substitution rate (i.e., 25%) of organic nitrogen using spent mushroom substrate effectively improves soil fertility, simultaneously optimizing photosynthetic capacity and root growth of grapevine seedlings. Full article

The article addresses the problem of optimising a selected production process in a company from the refractory products industry. As part of the research, individual activities were divided, identifying key wastes occurring in the production process. In addition, the 5S (the 5S methodology—Sort, Set in Order, Shine, Standardise, and Sustain) quality system was modified, its efficiency was increased, and a better work organisation was established based on it. Data from the actual production process were analysed based on total work efficiency using the OEE (Overall Equipment Effectiveness) coefficient. The use of machine working time was indicated, and key parameters were determined, i.e., availability, efficiency, and quality of the implemented production processes. The results obtained in the course of the research were compared to the Word Class OEE standards. The goal of the work is to indicate possibilities and recommendations for increasing production efficiency without increasing costs, thanks to actions reducing the number of production defects and optimal distribution of employees on the production line. The presented analyses can help assess the management processes of other manufacturing companies operating in this highly specialised manufacturing sector. At the same time, the research conclusions enable other entities to evaluate the implementation of the proposed solutions in practice without incurring unnecessary financial outlays on improving production processes. Full article

We consider the distributed stochastic gradient descent problem, where a main node distributes gradient calculations among n workers. By assigning tasks to all workers and waiting only for the k fastest ones, the main node can trade off the algorithm’s error with its runtime by gradually increasing k as the algorithm evolves. However, this strategy, referred to as adaptive k-sync, neglects the cost of unused computations and of communicating models to workers that reveal a straggling behavior. We propose a cost-efficient scheme that assigns tasks only to k workers, and gradually increases k. To learn which workers are the fastest while assigning gradient calculations, we introduce the use of a combinatorial multi-armed bandit model. Assuming workers have exponentially distributed response times with different means, we provide both empirical and theoretical guarantees on the regret of our strategy, i.e., the extra time spent learning the mean response times of the workers. Furthermore, we propose and analyze a strategy that is applicable to a large class of response time distributions. Compared to adaptive k-sync, our scheme achieves significantly lower errors with the same computational efforts and less downlink communication while being inferior in terms of speed. Full article

Wrist fractures, especially those involving the elbow and distal radius, are the most common injuries in children, teenagers, and young adults, with the highest occurrence rates during adolescence. However, the demand for medical imaging and the shortage of radiologists make it challenging to ensure accurate diagnosis and treatment. This study explores how AI-driven approaches are used to enhance fracture detection and improve diagnostic accuracy. In this paper, we propose the latest version of YOLO (i.e., YOLO11) with an attention module, designed to refine detection correctness. We integrated attention mechanisms, such as Global Attention Mechanism (GAM), channel attention, and spatial attention with Residual Network (ResNet), to enhance feature extraction. Moreover, we developed the ResNet_GAM model, which combines ResNet with GAM to improve feature learning and model performance. In this paper, we apply a data augmentation process to the publicly available GRAZPEDWRI-DX dataset, which is widely used for detecting radial bone fractures in X-ray images of children. Experimental findings indicate that integrating Squeeze-and-Excitation (SE_BLOCK) into YOLO11 significantly increases model efficiency. Our experimental results attain state-of-the-art performance, measured by the mean average precision (mAP50). Through extensive experiments, we found that our model achieved the highest mAP50 of 0.651. Meanwhile, YOLO11 with GAM and ResNet_GAM attained a maximum precision of 0.799 and a recall of 0.639 across all classes on the given dataset. The potential of these models to improve pediatric wrist imaging is significant, as they offer better detection accuracy while still being computationally efficient. Additionally, to help surgeons identify and diagnose fractures in patient wrist X-ray images, we provide a Fracture Detection Web-based Interface based on the result of the proposed method. This interface reduces the risk of misinterpretation and provides valuable information to assist in making surgical decisions. Full article

Background/Objectives: Developing a treatment strategy that effectively prolongs the lives of people with brain tumors requires an accurate diagnosis of the condition. Therefore, improving the preoperative classification of meningiomas is a priority. Machine learning (ML) has made great strides thanks to the development of convolutional neural networks (CNNs) and computer-aided tumor detection systems. The deep convolutional layers automatically extract important and dependable information from the input space, in contrast to more traditional neural network layers. One recent and promising advancement in this field is ML. Still, there is a dearth of studies being carried out in this area. Methods: Therefore, starting with the analysis of magnetic resonance images, we have suggested in this research work a tried-and-tested and methodical strategy for real-time meningioma diagnosis by image segmentation using a very deep transfer learning CNN model or DNN model (VGG-16) with CUDA. Since the VGGNet CNN model has a greater level of accuracy than other deep CNN models like AlexNet, GoogleNet, etc., we have chosen to employ it. The VGG network that we have constructed with very small convolutional filters consists of 13 convolutional layers and 3 fully connected layers. Our VGGNet model takes in an sMRI FLAIR image input. The VGG’s convolutional layers leverage a minimal receptive field, i.e., 3 × 3, the smallest possible size that still captures up/down and left/right. Moreover, there are also 1 × 1 convolution filters acting as a linear transformation of the input. This is followed by a ReLU unit. The convolution stride is fixed at 1 pixel to keep the spatial resolution preserved after convolution. All the hidden layers in our VGG network also use ReLU. A dataset consisting of 264 3D FLAIR sMRI image segments from three different classes (meningioma, tuberculoma, and normal) was employed. The number of epochs in the Sequential Model was set to 10. The Keras layers that we used were Dense, Dropout, Flatten, Batch Normalization, and ReLU. Results: According to the simulation findings, our suggested model successfully classified all of the data in the dataset used, with a 99.0% overall accuracy. The performance metrics of the implemented model and confusion matrix for tumor classification indicate the model’s high accuracy in brain tumor classification. Conclusions: The good outcomes demonstrate the possibility of our suggested method as a useful diagnostic tool, promoting better understanding, a prognostic tool for clinical outcomes, and an efficient brain tumor treatment planning tool. It was demonstrated that several performance metrics we computed using the confusion matrix of the previously used model were very good. Consequently, we think that the approach we have suggested is an important way to identify brain tumors. Full article

This article presents the simulation results of synchronization of a permanent magnet synchronous generator (PMSG) dedicated for a hydroelectric plant without power converter devices. The proposed machine design allows to connect a generator to the grid in two different ways. With the first method, the machine is connected to the grid in a similar way as in the case of an electrically excited synchronous generator. The second method is a direct line-start process based on asynchronous torque—similar to asynchronous motor start. Both methods can be used alternately. The advantages of the presented design are elimination of converter devices for starting the PMSG, possibility of use in small and medium hydroelectric power plants, operation with a high efficiency and high power factor in a wide range of generated power, and smaller dimensions in comparison to the generators currently used. The described rotor design allows for the elimination of capacitor batteries for compensation of reactive power drawn by induction generators commonly used in small hydroelectric plants. In addition, due to the high efficiency of the PMSG, high power factor, and appropriately selected design, the starting current during synchronization is smaller than in the case of an induction generator, which means that the structural elements wear out more slowly, and thus, the generator’s service life is increased. In this work, it is shown that PMSG with a rotor cage should have permanent magnets with an increased temperature class in order to avoid demagnetization of the magnets during asynchronous start-up. In addition, manufacturers of such generators should provide the number of start-up cycles from cold and warm states in order to avoid shortening the service life of the machine. The main objective of the article is to present the methods of synchronizing a generator of such a design (a rotor with permanent magnets and a starting cage) and their consequences on the behavior of the machine. The presented design allows synchronization of the generator with the network in two ways. The first method enables synchronization of the generator with the power system by asynchronous start-up, i.e., obtaining a starting torque exceeding the braking torque from the magnets. The second method of synchronization is similar to the method used in electromagnetically excited generators, i.e., before connecting, the rotor is accelerated to synchronous speed by means of a water turbine, and then, the machine is connected to the grid by switching on the circuit breaker. This paper presents electromagnetic phenomena occurring in both cases of synchronization and describes the influence of magnet temperature on physical quantities. Full article

This paper presents a method for dynamic pattern recognition and classification of one dangerous caterpillar species to allow for its control in maize crops. The use of dynamic pattern recognition supports the identification of patterns in digital image data that change over time. In fact, identifying fall armyworms (Spodoptera frugiperda) is critical in maize production, i.e., in all of its growth stages. For such pest control, traditional agricultural practices are still dependent on human visual effort, resulting in significant losses and negative impacts on maize production, food security, and the economy. Such a developed method is based on the integration of digital image processing, multivariate statistics, and machine learning techniques. We used a supervised machine learning algorithm that classifies data by finding an optimal hyperplane that maximizes the distance between each class of caterpillar with different lengths in N-dimensional spaces. Results show the method’s efficiency, effectiveness, and suitability to support decision making for this customized control context. Full article

Class-incremental learning (CIL) enables models to continuously acquire knowledge and adapt in an ever-changing environment. However, one primary challenge lies in the trade-off between the stability and plasticity, i.e., plastically expand the novel knowledge base and stably retaining previous knowledge without catastrophic forgetting. We find that even recent promising CIL methods via pre-trained models (PTMs) still suffer from this dilemma. To this end, this paper begins by analyzing the aforementioned dilemma from the perspective of marginal distribution for data categories. Then, we propose the feature transformation tuning (FeTT) model, which concurrently alleviates the inadequacy of previous PTM-based CIL in terms of stability and plasticity. Specifically, we apply the parameter-efficient fine-tuning (PEFT) strategies solely in the first CIL task to bridge the domain gap between the PTMs and downstream task dataset. Subsequently, the model is kept fixed to maintain stability and avoid discrepancies in training data distributions. Moreover, feature transformation is employed to regulate the backbone representations, boosting the model’s adaptability and plasticity without additional training or parameter costs. Extensive experimental results and further feature channel activations discussion on CIL benchmarks across six datasets validate the superior performance of our proposed method. Full article

Image segmentation is one of the key factors in diagnosing glioma patients with brain tumors. It helps doctors identify the types of tumor that a patient is carrying and will lead to a prognosis that will help save the lives of patients. The analysis of medical images is a specialized domain in computer vision and image processing. This process extracts meaningful information from medical images that helps in treatment planning and monitoring the condition of patients. Deep learning models like CNN have shown promising results in image segmentation by identifying complex patterns in the image data. These methods have also shown great results in tumor segmentation and the identification of anomalies, which assist health care professionals in treatment planning. Despite advancements made in the domain of deep learning for medical image segmentation, the precise segmentation of tumors remains challenging because of the complex structures of tumors across patients. Existing models, such as traditional U-Net- and SAM-based architectures, either lack efficiency in handling class-specific segmentation or require extensive computational resources. This study aims to bridge this gap by proposing Segment Anything Model 2-UNetwork, a hybrid model that leverages the strengths of both architectures to improve segmentation accuracy and consumes less computational resources by maintaining efficiency. The proposed model possesses the ability to perform explicitly well on scarce data, and we trained this model on the Brain Tumor Segmentation Challenge 2020 (BraTS) dataset. This architecture is inspired by U-Networks that are based on the encoder and decoder architecture. The Hiera pre-trained model is set as a backbone to this architecture to capture multi-scale features. Adapters are embedded into the encoder to achieve parameter-efficient fine-tuning. The dataset contains four channels of MRI scans of 369 glioma patients as T1, T1ce, T2, and T2-flair and a segmentation mask for each patient consisting of non-tumor (NT), necrotic and non-enhancing tumor (NCR/NET), and peritumoral edema or GD-enhancing tumor (ET) as the ground-truth value. These experiments yielded good segmentation performance and achieved balanced performance based on the metrics discussed next in this paragraph for each tumor region. Our experiments yielded the following results with minimal hardware resources, i.e., 16 GB RAM with 30 epochs: a mean Dice score (mDice) of 0.771, a mean Intersection over Union (mIoU) of 0.569, an $S_{\alpha }$ score of 0.692, a weighted F-beta score ($F_{\beta }^w$) of 0.267, a F-beta score ($F_{\beta }$) of 0.261, an $E_{\varphi }$ score of 0.857, and a Mean Absolute Error (MAE) of 0.04 on the BraTS 2020 dataset. Full article

This paper examines the changes in the performance level and pollutant emissions of a combustion chamber for turbofan engines. Two different fuels are compared: a conventional liquid fuel of the JET-A (kerosene) class and a hydrogen-based gaseous fuel. A turbofan engine delivering a 70 kN thrust at cruise conditions and 375 kN thrust at take-off is considered. The comparison is carried out by investigating the combustion pattern with different boundary conditions, the latter assigned along a typical flight mission. The calculations rely on a combined approach with a preliminary lumped parameter estimation of the engine performance and thermodynamic properties under different flight conditions (i.e., take-off, climbing, and cruise), and a CFD-based combustion simulation employing as boundary conditions the outputs obtained from the 0-D computations. The results are discussed in terms of performance, thermal properties, distributions throughout the combustor, and of pollutant concentration at the combustor outflow. The results demonstrate that replacing the JET-A fuel with hydrogen does not affect the overall engine performance significantly, and stable and efficient combustion takes place inside the burner, although a different temperature regime is observable causing a relevant increase in thermal NO emissions. Full article

Extracellular vesicles (EVs) and their cargo are increasingly suggested as innovative biomarkers correlated to the diagnosis, progression and therapy of diseases like cancer. Several techniques have been developed for the specific separation of the different classes of EVs that give solutions enriched in vesicles, but still containing other unwanted components. New methods for a more efficient, reliable and automated isolation of EVs are therefore highly desirable. Here, microparticles with surfaces endowed with positive ions were exploited to separate vesicles from complex biological matrices. First, flat silicon oxide surfaces functionalized with different divalent cations were tested for their efficiency in terms of small EV capture. Small EVs pre-purified via serial ultracentrifugations were employed for these analyses. The two better-performing cations, i.e., Cu²⁺ and Ni²⁺, were then selected to functionalize magnetic microbeads to be inserted in microfluidic chips and evaluated for their efficiency in capturing EVs and for their release of biomarkers. The best protocol setup was explored for the capture of EVs from cell culture supernatants and for the analysis of a class of biomarkers, i.e., microRNAs, via RT-PCR. The promising results obtained with this on-chip protocol evidenced the potential automation, miaturization, ease-of-use and the effective speed of the method, allowing a step forward toward its integration in simple and fast biosensors capable of analyzing the desired biomarkers present in EVs, helping the spread of biomarker analysis in both clinical settings and in research. Full article

Optical coherence tomography (OCT) is a cellular-resolution imaging technique that can be used as non-invasive and real-time imaging and is useful for detecting early stages of diseases. Five in vitro skin cells were measured by the Mirau-based full-field OCT, including keratinocyte (HaCaT cell line), melanocyte, squamous cell carcinoma cell line (A431), and two melanoma cell lines, i.e., A375 and A2058. Deep learning algorithms (particularly convolutional neural networks, CNN) that extract features from images efficiently process the OCT’s complex images. We used four models to classify the images of five types of 2D-OCT skin cells. Based on the ResNet-15 model, the mean accuracy (average accuracy of 10-fold cross-validation) reaches 98.47%, and the standard deviation is only 0.28% with the data augmentation method. Interestingly, while two normal skin cell images mix and the other three cancer skin cell images mix, the model still works to identify normal and cancer cell features. The mean accuracy reaches 96.77%. Furthermore, we used k-fold analysis to detect the model reliability and adopt the Gradient-weighted Class Activation Mapping (GRAD-CAM) to explain the discrimination results. The deep learning algorithm is successfully and efficiently applied to discriminate the OCT skin cell images. Full article

Remote sensing (RS) spectral time series provide a substantial source of information for the regular and cost-efficient monitoring of the Earth’s surface. Important monitoring tasks include land use and land cover classification, change detection, forest monitoring and crop type identification, among others. To develop accurate solutions for RS-based applications, often supervised shallow/deep learning algorithms are used. However, such approaches usually require fixed-length inputs and large labeled datasets. Unfortunately, RS images acquired by optical sensors are frequently degraded by aerosol contamination, clouds and cloud shadows, resulting in missing observations and irregular observation patterns. To address these issues, efforts have been made to implement frameworks that generate meaningful representations from the irregularly sampled data streams and alleviate the deficiencies of the data sources and supervised algorithms. Here, we propose a conceptually and computationally simple representation learning (RL) approach based on autoencoders (AEs) to generate discriminative features for crop type classification. The proposed methodology includes a set of single-layer AEs with a very limited number of neurons, each one trained with the mono-temporal spectral features of a small set of samples belonging to a class, resulting in a model capable of processing very large areas in a short computational time. Importantly, the developed approach remains flexible with respect to the availability of clear temporal observations. The signal derived from the ensemble of AEs is the reconstruction difference vector between input samples and their corresponding estimations, which are averaged over all cloud-/shadow-free temporal observations of a pixel location. This averaged reconstruction difference vector is the base for the representations and the subsequent classification. Experimental results show that the proposed extremely light-weight architecture indeed generates separable features for competitive performances in crop type classification, as distance metrics scores achieved with the derived representations significantly outperform those obtained with the initial data. Conventional classification models were trained and tested with representations generated from a widely used Sentinel-2 multi-spectral multi-temporal dataset, BreizhCrops. Our method achieved $77.06\%$ overall accuracy, which is $\sim 6\%$ higher than that achieved using original Sentinel-2 data within conventional classifiers and even $\sim 4\%$ better than complex deep models such as OmnisCNN. Compared to extremely complex and time-consuming models such as Transformer and long short-term memory (LSTM), only a $3\%$ reduction in overall accuracy was noted. Our method uses only $6.8k$ parameters, i.e., $\sim 400x$ fewer than OmnicsCNN and $\sim 27x$ fewer than Transformer. The results prove that our method is competitive in terms of classification performance compared with state-of-the-art methods while substantially reducing the computational load. Full article

The natural occurrence of harmful algal blooms (HABs) adversely affects the quality of clean and fresh water. They pose increased risks to human health, aquatic ecosystems, and water bodies. Continuous monitoring and appropriate measures must be taken to combat HABs. Deep learning models that utilize computer vision play a vital role in identifying and classifying harmful algal blooms in aquatic environments and water storage facilities. Inspecting algal blooms using conventional methods, such as algae detection under microscopes, is difficult, expensive, and time-consuming. Deep learning algorithms have shown a notable and remarkable performance in the image classification domain and its applications, including microscopic algae species classification and detection. In this study, we propose a deep learning-based approach for classifying microscopic images of algae using computer vision. This approach employs a convolutional neural network (CNN) model integrated with two additional blocks—squeeze and dense blocks—to determine the presence of algae, followed by adding Grad-CAM to the proposed model to ensure interpretability and transparency. We performed several experiments on our custom dataset of microscopic algae images. Data augmentation techniques were employed to increase the number of images in the dataset, whereas pre-processing techniques were implemented to elevate the overall data quality. Our proposed model was trained on 3200 images consisting of four classes. We also compared our proposed model with the other transfer learning models, i.e., ResNet50 and Vgg16. Our proposed model outperformed the other two deep learning models. The proposed model demonstrated 96.7% accuracy, while Resnet50, EfficientNetB0, and VGG16 showed accuracy of 85.0%, 92.96%, and 93.5%, respectively. The results of this research demonstrate the potential of deep learning-based approaches for algae classification. This deep learning-based algorithm can be deployed in real-time applications to classify and identify algae to ensure the quality of water reservoirs. Computer-assisted solutions are advantageous for tracking freshwater algal blooms. Using deep learning-based models to identify and classify algae species from microscopic images is a novel application in the AI community. Full article