AI News

RSS

Enhancing Security Scanning with Infrared Thermography and CNNs

This study in Nature explores the application of convolutional neural networks (CNNs) in classifying infrared (IR) images for concealed object detection in security scanning. Leveraging a ResNet-50 model and transfer learning, the researchers refined pre-processing techniques such as k-means and fuzzy-c clustering to improve classification accuracy.

23 Apr 2024

Automation of Structural Health Monitoring for Civil Infrastructures Using AI and ML

This review explores the critical role of image-processing technologies in structural health monitoring (SHM) for civil infrastructures. It highlights the integration of artificial intelligence (AI) and machine learning (ML) to enhance SHM automation and accuracy. Various imaging modalities, including drones, thermography, LiDAR, and satellite imagery, are discussed for damage detection, crack identification, and deformation monitoring.

23 Apr 2024

Deep Learning for Computer-Assisted Interventions in Cataract Surgery

This study explores the transformative impact of deep learning (DL) techniques on computer-assisted interventions and post-operative surgical video analysis, focusing on cataract surgery. By leveraging large-scale datasets and annotations, researchers developed DL-powered methodologies for surgical scene understanding and phase recognition.

23 Apr 2024

The MPDB Dataset: Exploring Driving Behavior with Multimodal Physiological Data

Researchers present the MPDB dataset, capturing physiological responses of 35 participants during a driving simulator experiment. Combining EEG, ECG, EMG, GSR, and eye-tracking data with driving behaviors, the dataset offers insights into human cognitive functions while driving. Detailed collection methods, storage structures, and validation procedures ensure the dataset's reliability and effectiveness in studying driver behavior, paving the way for advancements in traffic psychology and behavior modeling.

22 Apr 2024

LCEFormer: Enhancing Remote Sensing Image Dehazing

Researchers introduce LCEFormer, a novel approach for remote sensing image dehazing, integrating CNN-based local context enrichment with transformer networks. Experiments on multiple datasets demonstrate state-of-the-art performance, surpassing existing methods in hazy scene restoration.

22 Apr 2024

Automating Information Extraction in Criminal Investigations Using a Hybrid Classification Model

Researchers developed a hybrid classification model to autonomously extract crucial information from legal documents, achieving superior accuracy in comparison to baseline models. Leveraging GPT-3.5 and employing prompt strategies, the model demonstrated efficiency in extracting key information from murder verdicts, offering a promising tool to enhance investigative workflows and decision-making in criminal investigations.

22 Apr 2024

Carrot Quality Assessment Using Hyperspectral Imaging and Multivariate Analysis

Through hyperspectral imaging (HSI) and multivariate analysis, researchers accurately predicted pH and carotenoid content in carrots, crucial for nutritional assessment. Models utilizing partial least squares regression (PLSR) and least squares support vector machine (LS-SVM) showed LS-SVM's superiority in pH prediction and carotenoid content, offering a promising approach for internal quality evaluation in carrots.

22 Apr 2024

Advancements in Electrodes for Wearable Skin Devices

The article explores electrode design for wearable skin devices, crucial for health monitoring and human-machine interfaces. It discusses properties like flexibility and conductivity and proposes methods like structure modification and hybrid materials. Applications range from health monitoring to therapy and human-machine interfaces, emphasizing the need for innovative electrode design to enhance device performance and integration with AI for smarter functionalities.

20 Apr 2024

Modular Time Representation in the Suprachiasmatic Nucleus Using ML

Using advanced imaging techniques and machine learning algorithms, researchers investigated the suprachiasmatic nucleus (SCN), the central circadian pacemaker in mammals. They discovered that collective calcium signals within the SCN correlated strongly with time of day, with distinct functional subtypes of neurons forming spatially organized modules. These modules exhibited ripple-like patterns representing specific time features, offering new insights into the systemic design principles of the biological clock.

20 Apr 2024

OCTDL: An Open-Access OCT Dataset for Deep Learning in Ophthalmology

Researchers introduced OCTDL, an open-access dataset comprising over 2000 labeled OCT images of retinal diseases, including AMD, DME, and others. Utilizing high-resolution OCT scans obtained from an Optovue Avanti RTVue XR system, the dataset facilitated the development of deep learning models for disease classification. Validation with VGG16 and ResNet50 architectures demonstrated high performance, indicating OCTDL's potential for advancing automatic processing and early disease detection in ophthalmology.

20 Apr 2024

Automated Detection of Epiretinal Membranes in OCT Scans

Researchers developed a deep neural network (DNN) ensemble to automatically detect and classify epiretinal membranes (ERMs) in optical coherence tomography (OCT) scans of the macula. Leveraging over 11,000 images, the ensemble achieved high accuracy, particularly in identifying small ERMs, aided by techniques like mixup for data augmentation and t-stochastic neighborhood embeddings (t-SNE) for dimensional reduction.

20 Apr 2024

Traffic Coordination in VANETs with Probabilistic Manhattan Grid Topology

Researchers proposed a novel approach utilizing a modified Manhattan grid topology to address challenges in vehicular communication technology, focusing on India's dynamic traffic conditions. Their methodology integrated the fuzzy c-means (FCM) algorithm for jamming attacker detection, the modified fisheye state routing (MFSR) algorithm for efficient data exchange, and the particle swarm optimization (PSO) algorithm for precise attacker localization.

18 Apr 2024

AI-Driven Enhancement of Retinal Pigment Epithelial Cell Visualization Using Adaptive OCT

Researchers developed a novel AI method, P-GAN, to improve the visualization of retinal pigment epithelial (RPE) cells using adaptive optics optical coherence tomography (AO-OCT). By transforming single noisy images into detailed representations of RPE cells, this approach enhances contrast and reduces imaging time, potentially revolutionizing ophthalmic diagnostics and personalized treatment strategies for retinal conditions.

18 Apr 2024

AI-Enhanced Monitoring of Mediterranean Seagrass for Conservation and Climate Resilience

Researchers leverage deep-learning AI models and Sentinel-2 satellite data to map and monitor threatened Posidonia oceanica (P. oceanica) meadows in the Mediterranean. Their operational pipeline, including data acquisition, DL neural networks, and post-processing, achieves high accuracy. This approach supports conservation efforts, aiding long-term monitoring and informing decisions for seagrass preservation amid climate change.

18 Apr 2024

Advancements in Human Action Recognition: A Deep Learning Perspective

The paper explores human action recognition (HAR) methods, emphasizing the transition to deep learning (DL) and computer vision (CV). It discusses the evolution of techniques, including the significance of large datasets and the emergence of HARNet, a DL architecture merging recurrent and convolutional neural networks (CNN).

18 Apr 2024

Machine Learning-based System for Analyzing Sexual Harassment in Middle Eastern Literature

Scholars utilized machine learning techniques to analyze instances of sexual harassment in Middle Eastern literature, employing lexicon-based sentiment analysis and deep learning architectures. The study identified physical and non-physical harassment occurrences, highlighting their prevalence in Anglophone novels set in the region.

17 Apr 2024

Enhanced PSO Algorithm for Engineering Problems

Researchers introduced the NDWPSO algorithm, enhancing PSO's performance for engineering optimization. Integrating various strategies like dynamic inertia weight, adaptive acceleration coefficients, and innovative initialization schemes, NDWPSO demonstrated superior convergence speed and solution accuracy.

17 Apr 2024

Edge Computing LoRa Gateway for Real-Time Messaging

Researchers developed an edge-computing LoRaWAN gateway for real-time confirmed messaging, significantly reducing messaging time and network server resource usage. The gateway's innovative design and hardware components optimized performance and lowered costs, making it suitable for diverse IoT applications.

17 Apr 2024

Predicting Preplaced Aggregate Concrete Strength with Machine Learning

Researchers evaluated 13 machine learning models to forecast compressive strength in preplaced aggregate concrete. Extreme gradient boosting (XGBoost) emerged as the most accurate, with sensitivity and SHAP analyses highlighting crucial factors like gravel and water-to-binder ratio.

17 Apr 2024

Robotic Labs for Streamlining Material Synthesis with Thermodynamic Strategies

Researchers introduced a thermodynamic strategy for precursor selection to optimize synthesis pathways for complex materials, addressing challenges in phase purity. Utilizing a robotic laboratory, they validated this approach across 224 reactions, achieving higher phase purity for 35 target oxides.

16 Apr 2024