Transformer-based medication recommendation with a multiple graph augmentation strategy

Expert Systems with Applications, Год журнала: 2024, Номер 257, С. 125091 - 125091

Опубликована: Авг. 13, 2024

Язык: Английский

---

Machine reading comprehension based named entity recognition for medical text

Multimedia Tools and Applications, Год журнала: 2025, Номер unknown

Опубликована: Янв. 7, 2025

Язык: Английский

---

Human skin-inspired neuromorphic sensors

Soft Science, Год журнала: 2025, Номер 5(2)

Опубликована: Март 21, 2025

Human skin-inspired neuromorphic sensors have shown great potential in revolutionizing machines to perceive and interact with environments. skin is a remarkable organ, capable of detecting wide variety stimuli high sensitivity adaptability. To emulate these complex functions, been engineered flexible or stretchable materials sense pressure, temperature, texture, other physical chemical factors. When integrated computing systems, which the brain’s ability process sensory information efficiently, can further enable real-time, context-aware responses. This study summarizes state-of-the-art research on principles computing, exploring their synergetic create intelligent adaptive systems for robotics, healthcare, wearable technology. Additionally, we discuss challenges material/device development, system integration, computational frameworks human sensors, highlight promising directions future research.

Язык: Английский

---

A novel reconstruction-based video anomaly detection with idempotent generative network

Alexandria Engineering Journal, Год журнала: 2025, Номер 124, С. 513 - 525

Опубликована: Апрель 11, 2025

Язык: Английский

---

Multi-Object Tracking based on Optimal Transport and Coordinate Attention Mechanism

Signal Processing, Год журнала: 2025, Номер unknown, С. 110058 - 110058

Опубликована: Апрель 1, 2025

Язык: Английский

---

Automatic Detection and Classification of Natural Weld Defects Using Alternating Magneto-Optical Imaging and ResNet50

Sensors, Год журнала: 2024, Номер 24(23), С. 7649 - 7649

Опубликована: Ноя. 29, 2024

It is difficult to detect and identify natural defects in welded components. To solve this problem, according the Faraday magneto-optical (MO) effect, a nondestructive testing system for MO imaging, excited by an alternating magnetic field, established. For acquired images of crack, pit, lack penetration, gas pore, no defect, Gaussian filtering, bilateral median filtering are applied image preprocessing. The effectiveness these methods evaluated using metrics such as peak signal-noise ratio (PSNR) mean squared error. Principal component analysis (PCA) employed extract column vector features from downsampled defect images, which then serve input layer error backpropagation (BP) neural network model support machine (SVM) model. These two models can be used classification partial but recognition accuracy cracks pores comparatively low. further enhance weld defects, convolutional (CNN) ResNet50 established, parameters optimized. experimental results show that overall 99%. Compared with PCA-SVM CNN model, was increased 7.4% 1.8%, pore 10% 4%, respectively, indicating effectively accurately classify defects.

Язык: Английский

---

A forest fire detection method based on improved YOLOv5

Signal Image and Video Processing, Год журнала: 2024, Номер 19(2)

Опубликована: Дек. 16, 2024

Язык: Английский

---

Research on Classification and Identification of Crack Faults in Steam Turbine Blades Based on Supervised Contrastive Learning

Entropy, Год журнала: 2024, Номер 26(11), С. 956 - 956

Опубликована: Ноя. 6, 2024

Steam turbine blades may crack, break, or suffer other failures due to high temperatures, pressures, and high-speed rotation, which seriously threatens the safety reliability of equipment. The signal characteristics different fault types are slightly different, making it difficult accurately classify faults rotating directly through vibration signals. This method combines a one-dimensional convolutional neural network (1DCNN) channel attention mechanism (CAM). 1DCNN can effectively extract local features time series data, while CAM assigns weights each highlight key features. To further enhance efficacy feature extraction classification accuracy, projection head is introduced in this paper systematically map all sample into normalized space, thereby improving model's capacity distinguish between distinct types. Finally, optimization supervised contrastive learning (SCL) strategy, model better capture subtle differences Experimental results show that proposed has an accuracy 99.61%, 97.48%, 96.22% task multiple crack at three speeds, significantly than Multilayer Perceptron (MLP), Residual Network (ResNet), Momentum Contrast (MoCo), Transformer methods.

Язык: Английский

---

LMGA: Lightweight multi-graph augmentation networks for safe medication recommendation

Journal of King Saud University - Computer and Information Sciences, Год журнала: 2024, Номер 36(10), С. 102245 - 102245

Опубликована: Ноя. 17, 2024

Язык: Английский

---