Search

Abstract Accurate crack detection is crucial for maintaining pavement integrity, yet manual inspections remain labor-intensive and prone to errors, underscoring the need for automated solutions. This study proposes a novel crack segmentation approach utilizing advanced visual models, specifically Detectron2 and the Segment Anything Model (SAM), applied to the CFD and Crack500 datasets, which exhibit intricate and diverse crack patterns. Detectron2 was tested with four configurations—mask_rcnn_R_50_FPN_3x, mask_rcnn_R_101_FPN_3x, faster_rcnn_R_50_FPN_3x, and faster_rcnn_R_101_FPN_3x—while SAM was compared using Focal Loss, DiceCELoss, and DiceFocalLoss. SAM with DiceFocalLoss outperformed Detectron2, achieving mean IoU scores of 0.69 and 0.59 on the CFD and Crack500 datasets, respectively. The integration of Detectron2 with faster_rcnn_R_101_FPN_3x and SAM using DiceFocalLoss involves generating bounding boxes with Detectron2, which serve as prompts for SAM to produce segmentation masks. This approach achieves mIoU scores of 0.83 for CFD dataset and 0.75 for Crack500 dataset. These results highlight the potential of combining foundation models with Detectron2 for advancing crack detection technologies, offering valuable insights for enhancing highway maintenance systems.

Atomic layer deposition (ALD) has been identified as a promising growth method for high-uniformity superconducting thin films for superconducting quantum photonic applications, offering superior uniformity, thickness control and conformality to techniques such as reactive sputtering. The potential scalability of ALD makes this method especially appealing for fabrication of superconducting nanowires and resonators across large areas. We report on the growth of highly uniform superconducting NbN thin films via plasma-enhanced atomic layer deposition (PEALD) with radio frequency substrate biasing, on a 200 mm (8 inch) Si wafer, specifically for superconducting nanowire single-photon detector applications. Niobium nitride films were grown using (tert-butylimido)-tris(diethylamido)-niobium(V) precursor and an H _2 /Ar plasma. The superconducting properties of a variable thickness series of films (5.9–29.8 nm) show critical temperature ( T _c ) of 13.5 K approaching bulk thickness (28.8 nm) with low suppression down to the ultrathin regime (5.9 nm), with T _c = 10.2 K. T _c across the 200 mm wafer with 8 nm thick NbN, measured in 15 mm intervals, exhibits minimal variation ( 10 MA cm ^−2 at 2.6 K. PEALD could therefore be a pivotal technique in enabling large-scale fabrication of integrated quantum photonic devices across a variety of applications.

Abstract The study explores end-to-end deep learning frameworks and ensemble methods to enhance the accuracy of anatomical landmark identification in cephalometric radiographs, crucial for precise cephalometric analysis and effective orthodontic treatment planning. The methodology is strategically designed to address the complexities and variations in cephalometric images through a dual learning process: building base learners using end-to-end regression approaches and then constructing a meta-learner that ensembles the output of the base models to obtain collective predictions. Two primary approaches are followed to design base learners: an end-to-end single model and a multi-model strategy for localizing landmarks. Additionally, the efficacy of these approaches is examined by implementing two distinct deep architectures, namely UNet and UNetPlusPlus. The combination of the single model and multi-model approach with UNet and UNetPlusPlus results in four base learners, each displaying varying accuracy for different landmarks. To leverage the strengths of these individual models, three ensemble methods are explored, that includes supervised ensembling with XGBoost, unsupervised ensembling with the DBSCAN algorithm, and mathematical aggregation using weighted averaging. The methodology is evaluated on both public and private datasets, comprising 400 and 700 images, respectively. The multi-model approach demonstrates superior performance, achieving the highest detection accuracy of 79.9% (UNet) for the benchmark dataset and 92.8% (UNetPlusPlus) for the private dataset. The ensemble meta-model further boosts accuracy to 83.61% and 95.4%, respectively, reducing mean radial errors by 0.38 mm and 0.33 mm. These results highlight significant improvements in accuracy and error reduction through strategic combinations of deep learning architectures and ensemble techniques demonstrating the ability to significantly enhance cephalometric landmark annotation accuracy, which is critical for the practical applicability of the methodology.

Rapunzel syndrome is caused by ingestion of hairs, which leads to formation of trichobezoar, which can extend up to small intestine especially in girls. Utility of laparoscopic approach for this condition is not well established in this population. We are presenting a case of Rapunzel syndrome in a 14-year-old girl presenting as a cause of gastric outlet obstruction. Early diagnosis and extraction of trichobezoar by laparoscopic-assisted approach in children may be a better approach with respect to cosmesis, lesser morbidity, and early recovery.

Monitoring river water quality is crucial for safeguarding public health, protecting ecosystems, and ensuring economic sustainability. It helps detect contaminants, ensures drinking water safety, and facilitates early intervention for environmental protection and legal compliance. The objective of this study is to evaluate multiple machine learning algorithms to analyze water quality parameters in computing water quality index (WQI) and classification thereof, aiming to devise a reliable method for forecasting water quality with high accuracy. In this study, fourteen machine learning classifiers applied include Support Vector Machine (SVM), Random Forest (RF), Logistic Regression (LR), Decision Tree (DT), Multilayer Perceptron (MLP), K-Nearest Neighbor (KNN), Naïve Bayes, Gradient boosting, AdaBoost, Bagging, Extra Trees, Quadratic Discriminant Analysis (QDA), XGBoost, and CATBoost. A total of 1096 sample data was used where each data consists of nineteen analytical water quality parameters. To assess the performance of various classifiers, several evaluation techniques were utilized including confusion matrices, classification reports detailing precision and accuracy ratios, and Receiver Operating Characteristic (ROC) curves. The study also utilizes explainable AI (LIME and SHAP) to provide clear insights into the decision-making processes used to classify river water quality. The results indicated that all ML models demonstrate satisfactory performance in predicting WQI. Among the classifiers used, Gradient Boosting achieves the highest Accuracy (99.64 %), Precision (0.95), Recall (0.96), and F1-Score (0.95), indicating its superior ability to correctly classify instances and suggesting a balanced performance across different evaluation metrics. The analysis presented in this article holds the promise of providing accurate water quality data to researchers, thereby enhancing monitoring effectiveness through the application of machine learning techniques.

Background: Artificial intelligence (AI) is gaining prominence in the field of medicine with potential benefits in diagnosis, treatment, and medical education. However, the extent to which medical students are informed about AI, their attitudes toward the integration of AI with medicine, and their engagement in AI-related practices remains an essential aspect yet to be explored. Aims and Objectives: This study aims at assessing the knowledge, attitude, and practice of AI among medical students in a tertiary care hospital and understanding the perspectives of students toward AI as a teaching tool and in medical practice. Materials and Methods: This is a cross-sectional study using an online survey consisting of closed-ended questions under categories of demographical data, knowledge, attitude, practice, and discussion among medical students of a government medical college and hospital in Kallakurichi. Results: Among 150 participants of the study, 74.16% were found to have a basic knowledge of AI cumulatively. 49.33% of the participants reported that they expected a positive impact of AI in health care; 14.0% shared that AI could have a negative impact on existing healthcare practices; and the remaining 36.67% were impartial toward the outcome of AI integration. 10.67% of participants were extremely confident in their AI skills in medical practice; 26% reported that they were somewhat confident; 45.33% believed that they could adapt if needed; and 18% were uncertain of it. Finally, 17.33% were accepting toward AI evaluating their clinical skills, 36.0% were in strong denial of AI-based evaluation, and 46.67% agreed that it could be useful in some specific cases. Conclusion: In conclusion, the study highlights the current level of knowledge and attitudes of medical students toward AI. Incorporating AI tools and knowledge into medical education may lead to more informed and positively inclined future health-care professionals who can harness the potential benefits of AI in health care and advance medical knowledge. [ABSTRACT FROM AUTHOR]

Background: There is a prevalence of child sexual abuse between 2017-20 shows a result of 80% of the sufferers being females under the age of 14. In India, child abuse can take numerous forms. According to the National Commission for the Protection of Child Rights (NCPCR), over 55% of Indian children have experienced some form of abuse. This amount is extremely high and indicates an extremely serious issue that must be addressed. Physical abuse is one of the most common forms of child maltreatment in India, affecting approximately 25% of children at some point. The most prevalent varieties are slapping and hitting, as well as burning and pounding. Around 16% of children have experienced sexual abuse in some form, such as rape or unwanted sexual contact. Mental harassment is also on the rise, with 11% of youngsters reporting emotional or verbal abuse from caretakers. Objective: The main aim of this study, is to identify the antecedents and consequences child abuse among children between 7 to 17 years old with a structured questionnaire. Participants and Setting: 120 students between the ages 10 and 16 from South India, Tamil Nadu, Chennai government school in India. Methodology: The methodology used in this study has a descriptive approach and data was analyzed with a quantitative strategy. Results: There is a significant model fit among the dimensions of child abuse and neglect which leads to the consequences of mental, physical, and health problems. Conclusions: The purpose of exploring the child abuse is to lesser the possible chance of neglect or child abuse, and lower childhood trauma. There is a need to focus each state-wise by implementing comprehensive, evidencebased, community-driven child abuse and neglect prevention strategies based upon each and every state's needs. [ABSTRACT FROM AUTHOR]

Dielectric and electrocaloric (EC) properties of sol-gel synthesized BaTi0.84Zr0.16O3 (BTZ) and La0.78Sr.0.22MnO3 (LSM) composite in 80:20 molar ratio is reported. We discuss the dielectric analysis based on the electric modulus M*(=1/ϵ*) with the aim to differentiate the orientational and interfacial/carrier contribution and to explain their temperature and frequency dependence. Frequency dependence of imaginary part of M, i.e., M″(ω) of the composite exhibits a broad relaxation peak in MHz range while ϵ″(ω) graph is typical of interfacial/carrier polarization dominated: ϵ″ is as high as 7k at 20 Hz, decreases to 300 at 2 MHz measured at 25°C. M″-max shifts to higher frequencies with the increase in the temperature. The M″-peak, associated with BTZ-dipoles, is fitted using Cole-Cole (CC) equation to estimate the dielectric parameters, static and high-frequency dielectric constants ϵs and ϵꝏ respectively, relaxation time, τ, and the CC parameter, α, which is related to the distribution of relaxation time. Temperature dependence of physical parameters obtained from the fitting reflected the attributes of orientational polarization of the parent BTZ. These analyses indicate that M* is a more suitable parameter to consider for the ferroelectric-ferromagnetic composites.

Magneto-optical Kerr effect microscopy investigation of magnetocrystalline anisotropy and the magnetization reversal (MR) process in Co2FeSi (CFS) Heusler alloy thin films reveals the following: Regardless of the strength of the anti-site atomic disorder, all CFS films of fixed thickness (t = 50 nm) exhibit variations in the remanent magnetization (Mr) and coercive field (Hc) with the “in-plane” magnetic field (H) angle, φH, that are characteristic of the “in-plane” uniaxial anisotropy (UA) with an easy axis along φH = 0°. The observed variations with the field-angle are well described by the two-phase pinning (TP) model. Exceptions to this rule are the films with t ≥ 25 nm, having a maximum B2 atomic order, which show completely different angular variations in Mr and Hc. In such exceptional cases, we find that the TP model reproduces Mr(φH) and Hc(φH) only when it takes into account two mutually exclusive UAs, UA1 and UA2, with easy axes perpendicular to one another and UA1 ≫ UA2. When H points along the easy axis (φH = 0°), MR in all CFS films proceeds through the nucleation of reverse domains and their subsequent growth by domain wall movement as H increases. Atomic disorder has essentially no effect on the MR process, but the domain wall pinning at defects/imperfections affects the magnitudes of Hc and the UA field, Hk. At φH ≃ 45°, MR takes place through the formation of ripple domains. As H increases, the reverse domains grow at the expense of ripple domains, and a single domain configuration is established at high fields. When φH = 90°, MR in all films involves nucleation and subsequent growth of reverse domains by the field-induced movement of the 180° domain walls.

Objectives The objectives of this study are to explore and evaluate the automation of anatomical landmark localization in cephalometric images using machine learning techniques, with a focus on feature extraction and combinations, contextual analysis, and model interpretability through Shapley Additive exPlanations (SHAP) values. Methods We conducted extensive experimentation on a private dataset of 300 lateral cephalograms to thoroughly study the annotation results obtained using pixel feature descriptors including raw pixel, gradient magnitude, gradient direction, and histogram-oriented gradient (HOG) values. The study includes evaluation and comparison of these feature descriptions calculated at different contexts namely local, pyramid, and global. The feature descriptor obtained using individual combinations is used to discern between landmark and nonlandmark pixels using classification method. Additionally, this study addresses the opacity of LGBM ensemble tree models across landmarks, introducing SHAP values to enhance interpretability. Results The performance of feature combinations was assessed using metrics like mean radial error, standard deviation, success detection rate (SDR) (2 mm), and test time. Remarkably, among all the combinations explored, both the HOG and gradient direction operations demonstrated significant performance across all context combinations. At the contextual level, the global texture outperformed the others, although it came with the trade-off of increased test time. The HOG in the local context emerged as the top performer with an SDR of 75.84% compared to others. Conclusions The presented analysis enhances the understanding of the significance of different features and their combinations in the realm of landmark annotation but also paves the way for further exploration of landmark-specific feature combination methods, facilitated by explainability. [ABSTRACT FROM AUTHOR]

Hormone receptor positive (HR+) breast cancers are a heterogeneous class with differential prognosis. Although more than half of Indian women present with advanced disease, many such patients do well. We have attempted identification of biologically indolent tumors within HR+HER2- tumors based on gene expression using histological grade as a guide to tumor aggression. 144 HR+HER2- tumors were divided into subclasses based on scores derived by using transcript levels of multiple genes representing survival, proliferation, and apoptotic pathways and compared to classification by Ki-67 labeling index (LI). Clinical characters and disease free survival were compared between the subclasses. The findings were independently validated in the METABRIC data set. Using the previously established estrogen receptor (ER) down stream activity equation, 20% of the tumors with greater than 10% HR positivity by immunohistochemistry (IHC) were still found to have inadequate ER function. A tumor aggression probability score was used to segregate the remainder of tumors into indolent (22%) and aggressive (58%) classes. Significant difference in disease specific survival was seen between the groups (P = .02). Aggression probability based subclassification had a higher hazard ratio and also independent prognostic value (P