SkinNet-14: a deep learning framework for accurate skin cancer classification using low-resolution dermoscopy images with optimized training time

Neural Computing and Applications, Год журнала: 2024, Номер 36(30), С. 18935 - 18959

Опубликована: Авг. 1, 2024

Abstract The increasing incidence of skin cancer necessitates advancements in early detection methods, where deep learning can be beneficial. This study introduces SkinNet-14, a novel model designed to classify types using low-resolution dermoscopy images. Unlike existing models that require high-resolution images and extensive training times, SkinNet-14 leverages modified compact convolutional transformer (CCT) architecture effectively process 32 × pixel images, significantly reducing the computational load duration. framework employs several image preprocessing augmentation strategies enhance input quality balance dataset address class imbalances medical datasets. was tested on three distinct datasets—HAM10000, ISIC PAD—demonstrating high performance with accuracies 97.85%, 96.00% 98.14%, respectively, while time 2–8 s per epoch. Compared traditional transfer models, not only improves accuracy but also ensures stability even smaller sets. research addresses critical gap automated detection, specifically contexts limited resources, highlights capabilities transformer-based are efficient analysis.

Язык: Английский

---

Machine learning-based compressive strength estimation in nanomaterial-modified lightweight concrete

Open Engineering, Год журнала: 2024, Номер 14(1)

Опубликована: Янв. 1, 2024

Abstract The development of nanotechnology has led to the creation materials with unique properties, and in recent years, numerous attempts have been made include nanoparticles concrete an effort increase its performance create improved qualities. Nanomaterials are typically added lightweight (LWC) goal improving composite’s mechanical, microstructure, freshness, durability Compressive strength is most crucial mechanical characteristic for all varieties composites. For this reason, it essential accurate models estimating compressive (CS) LWC save time, energy, money. In addition, provides useful information planning construction schedule indicates when formwork should be removed. To predict CS mixtures or without nanomaterials, nine different were proposed study: gradient-boosted trees (GBT), random forest, tree ensemble, XGBoosted (XGB), Keras, simple regression, probabilistic neural networks, multilayer perceptron, linear relationship model. A total 2,568 samples gathered examined. significant factors influencing during modeling process taken into account as input variables, including amount cement, water-to-binder ratio, density, content aggregates, type nano, fine coarse aggregate content, water. suggested was assessed using a variety statistical measures, coefficient determination ( R 2 ), scatter index, mean absolute error, root-mean-squared error (RMSE). findings showed that, comparison other models, GBT model outperformed others predicting compression enhanced nanomaterials. produced best results, greatest value (0.9) lowest RMSE (5.286). Furthermore, sensitivity analysis that important factor prediction water content.

Язык: Английский

---

Belt Conveyor Idlers Fault Detection Using Acoustic Analysis and Deep Learning Algorithm with the YAMNet Pretrained Network

IEEE Sensors Journal, Год журнала: 2024, Номер 24(19), С. 31379 - 31394

Опубликована: Авг. 16, 2024

Язык: Английский

---

Advancing multiple sclerosis diagnosis through an innovative hybrid AI framework incorporating Multi-view ResNet and quantum RIME-inspired metaheuristics

Ain Shams Engineering Journal, Год журнала: 2025, Номер 16(2), С. 103241 - 103241

Опубликована: Янв. 13, 2025

Язык: Английский

---

Effect of Training Data Ratio and Normalizing on Fatigue Lifetime Prediction of Aluminum Alloys with Machine Learning

International Journal of Engineering, Год журнала: 2024, Номер 37(7), С. 1296 - 1305

Опубликована: Янв. 1, 2024

It is critical to evaluate the estimation of fatigue lifetimes for piston aluminum alloys, particularly in automotive industry. This paper investigates effect different normalization methods on performance lifetime using Extreme Gradient Boosting (XGBoost), as a supervised machine learning method. For this purpose, dataset used study includes various physical and experimental inputs related an alloy corresponding outputs. Furthermore, before fitting XGBoost model, preprocessing were utilized evaluated metrics such Root Mean Square Error (RMSE), Determination Coefficient (R2), Scatter Band (SB). The results indicate that modeling with logarithmic values method excels when trained 100% data. However, other demonstrate superior accuracy estimating test data 20% 80% train set split.

Язык: Английский

---

Navigating the Complexity of Money Laundering: Anti–money Laundering Advancements with AI/ML Insights

International Journal on Smart Sensing and Intelligent Systems, Год журнала: 2024, Номер 17(1)

Опубликована: Апрель 1, 2024

Abstract This study explores the fusion of artificial intelligence (AI) and machine learning (ML) methods within anti–money laundering (AML) frameworks using data from US Treasury’s Financial Crimes Enforcement Network (FinCEN). ML deep (DL) algorithms—such as random forest classifier, elastic net regressor, least absolute shrinkage selection operator (LASSO) regression, gradient boosting linear multilayer perceptron (MLP) convolutional neural network (CNN), K-nearest neighbor (KNN)—were used to forecast variables such state, year, transaction types (credit card debit card). Hyperparameter tuning through grid search randomized was optimize model performance. The results demonstrated efficacy AI/ML algorithms in predicting temporal, spatial, industry-specific money-laundering patterns. classifier achieved 99.99% average accuracy state prediction, while regressor excelled year simultaneously, credit transactions, respectively. MLP CNN showed promise context transactions. performed competitively with low mean squared error (MSE) (2.9) highest R -squared ( 2 ) value 0.24, showcasing its pattern-capturing proficiency. Logistic regression well area under receiver operating characteristic curve (ROC_AUC) scores 0.55 0.53, For a precision recall 0.42, 0.6 0.54, highlighting their effectiveness. recommends interpretability, hyperparameter optimization, specialized models, ensemble methods, augmentation, real-time monitoring for improved adaptability evolving financial crime Future improvements could include exploring integration blockchain technology AML.

Язык: Английский

---

Artificial intelligence for computer aided detection of pneumoconiosis: A succinct review since 1974

Engineering Applications of Artificial Intelligence, Год журнала: 2024, Номер 133, С. 108516 - 108516

Опубликована: Май 2, 2024

Язык: Английский

---

Automatic Recognition Technology of Library Books Based on Convolutional Neural Network Model

HighTech and Innovation Journal, Год журнала: 2024, Номер 5(1)

Опубликована: Март 1, 2024

Background: The development of technological devices has changed many facets our lives, particularly the way we engage with information and learning. advent automated technology for identification had a revolutionary effect on how read organize books within context data searches. It starts by solving difficulties in analyzing photos book pages using methods such as distortion rectification separation. Objective: research compares effectiveness suggested method traditional straight-line techniques real-world testing. Methodology: Skip-Gram model Word2Vec is used to accurately represent spoken language, allowing word vectors be generated input preprocessed CNN. results show that methodology created regarding present investigation works better than alternatives concerning accuracy efficiency during line identification. Result:This work advances field suggestion systems presenting strong effective leverages CNNs. findings demonstrate deep learning may optimize system recommendations improve customer service happiness variety contexts. This technique creates bridge between natural language processing picture evaluation opens up new possibilities advancement along user satisfaction. Doi: 10.28991/HIJ-2024-05-01-015 Full Text: PDF

Язык: Английский

---

NPGCL: neighbor enhancement and embedding perturbation with graph contrastive learning for recommendation

Applied Intelligence, Год журнала: 2025, Номер 55(6)

Опубликована: Фев. 5, 2025

Язык: Английский

---

High-order graph convolutional networks for circular Ribonucleic Acid and disease association prediction incorporating multiple biological relationships

Engineering Applications of Artificial Intelligence, Год журнала: 2025, Номер 146, С. 110303 - 110303

Опубликована: Фев. 21, 2025

Язык: Английский

---