2579 papers
This study proposes an improved Stacking ensemble learning algorithm that utilizes functional near-infrared spectroscopy (fNIRS) signals to achieve highly accurate, objective, and quantitative assessment of tinnitus severity, significantly outperforming conventional models and offering a promising tool for clinical diagnosis.
This paper proposes PixelChain Encryption (PCE), a lightweight, reversible cryptographic framework inspired by blockchain principles that secures images through recursive SHA-256 hashing and XOR-based keystreams without distributed ledgers, achieving high resistance to statistical attacks and low computational complexity suitable for resource-constrained IoT and medical imaging applications.
This paper proposes a novel image classification method for English letters using spiking neural P systems with structural plasticity and spike-timing-dependent plasticity, which combines gradient descent and Hebbian rules to achieve 99.79% accuracy on noise-free data and outperform existing models on noisy datasets.
This paper introduces CareLoop, a clinical-world sandbox that evaluates AI models on their ability to practice medicine within the complex context of patients' lives by measuring their performance in discovering hidden states, adapting to constraints, and managing consequences across simulated trajectories, revealing that execution-grounded capabilities are distinct from and more challenging than simply knowing medical facts.
This systematic review of 50 studies evaluates classical and deep learning segmentation techniques for malaria parasites and sickle cells in peripheral blood smears, highlighting their potential for field deployment in low-resource settings while identifying critical gaps in dataset standardization, external validation, and edge-device optimization.
This study evaluates vision-language models for extracting structured data from heterogeneous multi-page laboratory reports, finding that while whole-document processing offers superior speed, a page-by-page workflow using Qwen2.5-VL achieves the highest accuracy and stability across varying document lengths.
The paper introduces GLeMM, a large-scale, fully automated, and multilingual derivational morphology dataset covering seven European languages, designed to enable data-driven research and computational testing of form-meaning relations in word formation.
This paper proposes SOFA, a novel resource orchestration framework for geo-distributed Function-as-a-Service (FaaS) datacenters that simultaneously optimizes energy costs, carbon emissions, and water usage while maintaining or improving service-level agreement (SLA) performance.
This paper proposes a novel cross-modal framework for robust 3D representation learning that integrates a multi-granularity Gaussian mixture model for hierarchical probabilistic geometric modeling with a multi-scale visual enhancement module to achieve state-of-the-art performance in classification, part segmentation, and few-shot learning.
This paper presents the design and implementation of a prototype smart door lock that integrates YOLOv8n-based face recognition, a smart doorbell, and remote access control via Blynk and Supabase, demonstrating high accuracy and reliability in authenticating residents and managing visitor access for residential security.