Rotator cuff tears (RCTs) represent a common orthopedic condition, the diagnosis of which often requires advanced imaging techniques. Notably, plain radiography may not accurately differentiate between small to medium RCTs and other shoulder pathologies. Deep learning (DL) models may facilitate preliminary assessment of RCTs, helping avoid unnecessary advanced imaging and expediting patient care. For this study, we used a dataset comprising 587 shoulder radiographs (339 from patients with small to medium RCTs and 248 from without RCTs). The study dataset was divided into a training set (406 images [69
This study stress-tested the ELK Stack on three k3s clusters (Small Cluster, Medium Cluster, and Large Cluster VMs) on the TWCC platform using three configuration strategies. Results revealed a strong linear correlation ($R^{2}=0.983$) between throughput and the total number of CPU cores, with an average gain of approximately 4,169 events/sec per core. Performance strategies showed that the Config configuration balanced stability and performance in Small Cluster scenarios, while the Book configuration excelled in achieving stable, high output in Large Cluster environments. Anomalous drops in throughput were observed in the Basic 8c (part of the Medium Cluster) test, likely caused by transient resource contention in the cloud VM environment. The findings provide practical guidance for resource planning and optimal configuration selection for ELK deployment on cloud infrastructure.
Given the popularity of angling as a leisure activity in coastal countries, this study examined how different dimensions of anglers' activity involvement (attraction, centrality, and self-expression) affect psychological ownership and influence information-sharing behavior. Data were collected from 424 active anglers at the Northern Breakwater of the Taichung Port Sea Fishing Demonstration Area in Taiwan using systematic sampling. The findings reveal that the three dimensions of activity involvement activate different motivational pathways toward information-sharing behavior: attraction operates primarily through direct enjoyment-driven sharing; centrality functions exclusively through psychological ownership; and self-expression operates through both direct and ownership-mediated routes. These findings are most applicable to regulated, place-based recreational fishing settings, offering managers actionable insights on channeling anglers' general willingness to share toward governance-relevant content.
In underwater internet of things (UIoT) applications, the adoption of underwater image enhancement (UIE) technology can increase the image resolution, facilitating efficient visual exploration of underwater environments. However, low-resolution underwater images that exhibit blurring, low contrast, and color distortion make the construction of high-accuracy underwater vision systems based on UIoT architectures challenging. To address these problems, an energy-adaptive learning network inspired by nuclear fusion, namely, NCFDNet, is proposed. The network design of NCFDNet is a network design core that is based on simulated nuclear fusion physics heuristics for feature transfer and fusion. Multiscale fusion is achieved through this feature search core to reconstruct high-resolution feature information. NCFDNet is divided into three modules: a nuclear-core fusion-inspired energy-adaptive module, a color enhancement module, and a self-evolving multiscale high-resolution context enhancement module. The nuclear-core fusion-inspired energy-adaptive module relies on the concept of particle motion in nuclear physics to simulate the evolution mechanism of the feature weights in the deep learning network, and the color enhancement module dynamically enhances the RGB channels and performs multicolor space enhancement and fusion functions. Finally, the self-evolving multiscale high-resolution context enhancement module transfers these multiscale features and fuses them with the contextual features to reconstruct high-resolution underwater images.
