Browse Articles

Correction: The impact of floating photovoltaic power plants on lake water temperature and stratification

Plasmonic artificial inspector for herbal medicines via surface-enhanced Raman spectroscopy and deep learning

Human gut M cells resemble dendritic cells and present gluten antigen

The evolution of object detection from CNNs to transformers and multi-modal fusion

Abstract Object detection, a cornerstone of computer vision, aims to localize and classify objects within images. This comprehensive survey reviews modern object detection methods, focusing on two dominant paradigms: Convolutional Neural Networks (CNNs) and Transformer-based architectures. This work provides a structured comparison of CNN-based and Transformer-based detection paradigms, highlighting their complementary strengths and trade-offs. CNNs demonstrate advantages in local feature extraction and computational efficiency, whereas Transformers excel at capturing global context through self-attention mechanisms. We also analyze multi-modal fusion techniques integrating Red-Green-Blue (RGB), Light Detection and Ranging (LiDAR), and language embeddings. Benchmark results from representative models include: Real-Time Detection Transformer (RT-DETR) achieves 53.1% mean Average Precision (mAP) at Intersection over Union (IoU) at 0.5 : 0.95, You Only Look Once version 8 (YOLOv8) achieves 50.2% mAP at 0.5:0.95, real-time detectors exceed 100 frames per second (FPS) with competitive accuracy, and specialized infrared methods achieve 92.45% F-measure on NUAA-SIRST dataset. The work introduces a novel taxonomy of multi-modal fusion strategies, documents field-wide and review-specific limitations, and synthesizes recent 2024 to 2025 benchmarks across diverse datasets. Despite these advances, significant challenges remain in handling scale variation, occlusion effects, and domain adaptation. This survey outlines these persistent obstacles and promising research directions, providing a structured reference for researchers and practitioners.

Assessing the impact of infrastructure proliferation on shoreline dynamics around Mexico

Does AI already have human-level intelligence? The evidence is clear

Perillyl alcohol attenuates hypoxia induced right ventricular dysfunction and remodeling by balancing the renin angiotensin aldosterone system in rats

Are health influencers making us sick?

Efficient computation and design of high speed double precision Vedic multiplier architecture

Don’t talk science, play science: translate your data into music to improve its reach

Agonists for cytosolic bacterial receptor ALPK1 induce antitumour immunity

Combined fortification of yogurt with grape pomace pectic oligosaccharides and encapsulated probiotics

Psychosocial experience of couples coping with prostate cancer: a qualitative study

A network analysis of the associations between COVID-19-related variables and health across sex, age and educational levels among Ghanaian youths

A hybrid blockchain migration framework for converting traditional databases into blockchain-based EMR systems

Abstract Electronic Medical Records (EMRs) are crucial to modern healthcare. However, traditional relational databases fail to fulfill increased expectations for integrity, auditability, and compliance in regulated environments. This paper proposes a Hybrid Blockchain Migration Framework that integrates a conventional MySQL-based EMR system (OpenMRS) with a permissioned blockchain network (Hyperledger Fabric). Sensitive data fields are selectively mirrored to the blockchain, ensuring tamper-evident logging while retaining the high performance of SQL for routine operations. A middleware layer, implemented using Java Spring Boot, monitors changes in the EMR and commits cryptographic hashes and metadata to the blockchain in near real-time. We evaluate the hybrid system against both standalone MySQL and full-blockchain implementations using controlled benchmarks, analyzing latency, throughput, resource utilization, and auditability. Results show that the hybrid architecture sustains near-native responsiveness (median 2.1 ms versus 1.6 ms for pure MySQL and 60.5 ms for Fabric) and delivers 480 Transaction Per Second (TPS), while incurring only modest overhead (47% of i7-9750H CPU, 1.15 GB RAM) and enhancing data integrity and compliance with regulations such as Oman’s Personal Data Protection Law (PDPL). The framework is extensible to multi-institutional deployments and supports regulatory alignment, making it a viable pathway for blockchain adoption in clinical settings.

Utilising artificial intelligence to identify surgical anatomy during laparoscopic donor nephrectomy – a validation and feasibility study

Abstract Although the risk of intraoperative complications of laparoscopic donor nephrectomy (LDN) is now acceptably low, the work continues to minimise technical mishaps during this ‘high stakes’ surgery. In this study, we aim to demonstrate the pilot use of a patented proprietary deep learning (DL)-based computer vision (CV) to automatically recognise key anatomical structures and prevent intraoperative injuries, which is especially crucial during the learning curve. 6828 images manually annotated by pixels were selected from 16 surgical videos (National University Hospital, NUH) for training as ground truth, and 1757 annotated images from 4 separate surgical videos were used for validation. This ensured a balanced validation ratio of nearly 20% for each label (spleen, left kidney, renal artery, renal vein, and ureter). The YOLO (you only look once) v11x DL network ( https://docs.ultralytics.com/models/yolo11/ ), known for its speed and accuracy in real-time detection, was adapted to train our model. For further optimisation, it uses a sophisticated loss function which incorporates the accuracy of each pixel in segmentation tasks (binary cross-entropy loss), compares the predicted bounding box coordinates against ground truth (bounding box loss), and emphasises the importance of difficult-to-detect labels (distribution focal loss). Metrics were calculated using the following formulas, based on true positives (TP), false positives (FP), and false negatives (FN): Precision: TP / (TP + FP), Recall: TP / (TP + FN), F1 Score: 2 * (Precision * Recall) / (Precision + Recall). High precision minimises false positives, which could disrupt surgical workflows, while high recall ensures comprehensive detection, minimising false negatives that could affect patient safety. F1 serves as the harmonic mean of recall and precision. Quantitative evaluation of the validation dataset using the hold-out validation method yielded promising performance metrics and prospective evaluation was performed on a video from another surgeon (JC) and institution (National Taiwan University) and also in real-time in NUH. Our pilot study demonstrates an innovative machine learning design’s ability to accurately identify vital anatomical structures in LDN. This is a crucial first step for further artificial intelligence (AI)-guided applications such as intra-operative guidance, education, and post-hoc operative analysis and operative standards evaluation.

Largest galaxy survey yet confirms that the Universe is not clumpy enough

Enhancing corrosion resistance of Epoxy-Based composite coatings on mild steel using functionalized aluminium oxide (Al₂O₃) nanoparticles

Light-powered bacteria become living chemical factories

Comprehensive enhancement of PVC nanocomposites through Al2O3 for advanced optoelectronics

Abstract With the rapid advancement of elastic and translucent optoelectronics, the demand for low-cost materials has emerged as a major research topic. As a result, this paper investigated how Al 2 O 3 incorporation altered the properties of PVC. New PVC/xAl 2 O 3 (x = 0, 0.01, 0.03, 0.07, 0.1, 0.2 wt%) nanocomposites were manufactured by using a cost-effective and simple process (casting method) to adjust the Al 2 O 3 concentration. The PVC/xAl 2 O 3 (x = 0, 0.01, 0.03, 0.07, 0.1, 0.2 wt%) nanocomposites were analyzed using X-ray diffraction (XRD), Fourier Transform Infrared spectroscopy (FTIR), scanning electron microscopy (SEM), ultraviolet-visible (UV–visible) spectroscopy, and impedance measurements. Dislocation density (δ), Distortion parameter (g), nanoparticle size (D), and lattice strain (ε) were assessed using the Scherrer and Williamson–Hall methods. Crystal size and the number of crystallites were observed to increase with Al 2 O 3 content, revealing the higher crystallinity in the material. It was found that the particle size was ~ 30.22 nm for PVC/xAl 2 O 3 (x = 0.1wt%). SEM analysis showed a consistent distribution of Al 2 O 3 within the PVC at low concentrations of Al 2 O 3 . These PVC/Al 2 O 3 films were used as adjustable light- diffusing films in the packaging of different flexible photoelectric devices, according to their visible absorbance characteristic depending on filler concentrations. The optical bandgaps of PVC and PVC/x(Al 2 O 3 ) (where x = 0.2) were 5.05 eV and 3 eV, respectively. This decrease was associated with the creation of localized states in the bandgap. The refractive index values obtained were greater than those found in earlier studies, suggesting that the incorporation of a small amount of Al 2 O 3 nanoparticles improved the refractive index of PVC. It was noted that as the concentration of Al 2 O 3 rose, the dispersion parameters E d , M 2, and M 3 showed an increase, while E 0 exhibited a decrease. Conversely, the dielectric characteristics of the synthesized nanocomposites improved as the alumina content in the PVC matrix increased. The findings concluded that inexpensive PVC/xAl 2 O 3 (x = 0.1wt%) nanocomposites can be used as an essential component in sophisticated optoelectronic applications.