neural networks News and Research

RSS

UAVs and Computer Vision Enhance Remote Runway Inspections

Researchers developed an automated system utilizing UAVs and deep learning to monitor and maintain remote gravel runways in Northern Canada. This system accurately detects defects and evaluates runway smoothness, proving more effective and reliable than traditional manual methods in harsh and isolated environments.

5 Jun 2024

Permafrost Embankment Thermal Prediction via Deep Learning

A novel approach integrates deep learning with geotechnical knowledge to predict the stochastic thermal regime of permafrost embankments. Validated against real data, this method enhances accuracy and reduces computational costs, proving effective for diverse environmental conditions.

4 Jun 2024

Computer Vision and Deep Learning Enhance Emotion Recognition

A systematic review in the journal Sensors analyzed 77 studies on facial and pose emotion recognition using deep learning, highlighting methods like CNNs and Vision Transformers. The review examined trends, datasets, and applications, providing insights into state-of-the-art techniques and their effectiveness in psychology, healthcare, and entertainment.

4 Jun 2024

Computer Vision for Object Tracking: A Review

A comprehensive review highlights the evolution of object-tracking methods, sensors, and datasets in computer vision, guiding developers in selecting optimal tools for diverse applications.

4 Jun 2024

CNNs Optimize Waste Management in Smart Cities

Researchers evaluated deep learning models for waste classification in smart cities, with ResNeXt-101 emerging as the top performer. The study suggests a federated learning framework to enhance trash detection across diverse environments, leveraging multiple CNN models for improved efficiency in waste management.

3 Jun 2024

AI Enhances Pedestrian Monitoring in Train Stations

Researchers employed AI and computational vision techniques to improve pedestrian monitoring in crowded train stations. Utilizing YOLOv7 for object detection and AlphaPose for activity recognition, the study successfully tracked passenger movements and activities, providing critical insights for enhancing station safety and efficiency.

31 May 2024

CIMNet Advances Crop Disease Detection

Researchers introduced CIMNet, a novel network for crop disease image recognition, excelling in noisy environments. Featuring a non-local attention module and multi-scale critical information fusion, CIMNet outperformed traditional models in accuracy and applicability, significantly enhancing crop disease detection and improving agricultural productivity.

31 May 2024

Advancing PD Localization in Power Transformers Using ML and DL

This article in Scientific Reports compares ML and DL methods for localizing PD sources within power transformer tanks using single-sensor electric field measurements. Various techniques including CNN, SVR, SVM, BPNN, KNN, MLP, and XGBoost were evaluated across multiple case studies, demonstrating the CNN model's superior accuracy and robustness.

30 May 2024

Advanced Model Boosts Vessel Trajectory Classification

Researchers introduced biSAMNet, a cutting-edge model integrating word embedding and deep neural networks, for classifying vessel trajectories. Tested in the Taiwan Strait, it significantly outperformed other models, enhancing maritime safety and traffic management.

29 May 2024

Machine Learning Models for Turbulent Combustion Speed Prediction

Researchers developed and validated machine learning models for predicting turbulent combustion speed in hydrogen-natural gas spark ignition engines, showcasing their superiority over traditional methods. By leveraging data from a MINSEL 380 engine and employing techniques like random forest and artificial neural networks, the study demonstrated high forecasting accuracy, making these models valuable for industrial applications such as engine monitoring and simulation tools.

23 May 2024

Enhancing UAV Safety with Machine Learning

A study in Heliyon introduced a machine learning-based approach for predicting defects in BLDC motors used in UAVs. Researchers compared KNN, SVM, and Bayesian network models, with SVM demonstrating superior accuracy in fault classification, highlighting its potential for improving UAV operational safety and predictive maintenance.

23 May 2024

Seismic Damage Assessment Using Machine Learning

This article showcases a machine learning approach using K-nearest neighbors (KNN) and linear regression to assess seismic damage in moment-resisting frame buildings. By training on data generated through a new simulation procedure, researchers achieved accurate predictions of the Park-Ang structural damage index, with KNN demonstrating superior performance.

23 May 2024

LLMs Transform Drug Discovery with Molecule Translation

This study demonstrated the potential of T5 large language models (LLMs) to translate between drug molecules and their indications, aiming to streamline drug discovery and enhance treatment options. Using datasets from ChEMBL and DrugBank, the research showcased initial success, particularly with larger models, while identifying areas for future improvement to optimize AI's role in medicine.

22 May 2024

Machine Learning Revolutionizes Plant Disease Detection

A recent article in "Artificial Intelligence in Agriculture" reviewed machine learning (ML) techniques for detecting plant diseases in apple, cassava, cotton, and potato crops. The study highlighted the superior accuracy of convolutional neural networks (CNNs) and emphasized ML's potential to enhance crop yield and quality, despite challenges related to data quality and ethical considerations.

22 May 2024

AI Unveils Key Turbulence Structures

Researchers used eXplainable AI (XAI) to identify critical coherent structures in wall-bounded turbulence, improving predictions of flow states. This novel approach, applicable to complex and high Reynolds number flows, enhances understanding and control of turbulent phenomena in engineering and natural systems.

21 May 2024

Enhanced CAD Integration Using Deep Learning

Researchers developed advanced deep learning (DL)-based automatic feature recognition (AFR) methods that significantly enhance computer-aided design (CAD), process planning (CAPP), and manufacturing (CAM) integration. Their approach, using the multidimensional attributed face-edge graph (maFEG) and Sheet-metalNet, a graph neural network, improves recognition accuracy and adapts to evolving datasets, addressing limitations of traditional and voxelized representations.

19 May 2024

Holographic System Blends Reality and Digital Content

Researchers present a groundbreaking holographic system in Nature, merging metasurface gratings, compact waveguides, and AI-driven holography algorithms to create vibrant 3D AR experiences. Their prototype, integrating a metasurface waveguide and phase-only SLM, achieves unmatched visual quality and represents a significant leap in wearable AR device development.

17 May 2024

Crack the Code: Advanced Method Enhances Train Rivet Inspection

This study proposes an innovative method for detecting cracks in train rivets using fluorescent magnetic particle detection (FMPFD) and instance segmentation, achieving high accuracy and recall. By enhancing the YOLOv5 algorithm and developing a single coil non-contact magnetization device, the researchers achieved significant improvements in crack detection.

17 May 2024

Mapping ML in Seismic Engineering

A recent scientometric review highlighted the transformative impact of machine learning (ML) in seismic engineering, showcasing advancements in material performance prediction and seismic resistance. The study, published in the journal Buildings, analyzed 3189 papers using the Scopus database, identifying key research trends and fostering collaboration within the field.

16 May 2024

DPA-1: A Deep Potential Model for Molecular Simulation

Researchers introduced DPA-1, a deep potential model with a gated attention mechanism, for representing atomic system conformation and chemical spaces. DPA-1 demonstrated superior performance in learning potential energy surfaces (PES) compared to existing benchmarks, offering efficiency and interpretability.

15 May 2024