Search

Abstract Background Accurate measurement of the hip-knee-ankle (HKA) angle is essential for informed clinical decision-making in the management of knee osteoarthritis (OA). Knee OA is commonly associated with varus deformity, where the alignment of the knee shifts medially, leading to increased stress and deterioration of the medial compartment. The HKA angle, which quantifies this alignment, is a critical indicator of the severity of varus deformity and helps guide treatment strategies, including corrective surgeries. Current manual methods are labor-intensive, time-consuming, and prone to inter-observer variability. Developing an automated model for HKA angle measurement is challenging due to the elaborate process of generating handcrafted anatomical landmarks, which is more labor-intensive than the actual measurement. This study aims to develop a ResNet-based deep learning model that predicts the HKA angle without requiring explicit anatomical landmark annotations and to assess its accuracy and efficiency compared to conventional manual methods. Methods We developed a deep learning model based on the variants of the ResNet architecture to process lower limb radiographs and predict HKA angles without explicit landmark annotations. The classification performance for the four stages of varus deformity (stage I: 0°–10°, stage II: 10°–20°, stage III: > 20°, others: genu valgum or normal alignment) was also evaluated. The model was trained and validated using a retrospective cohort of 300 knee OA patients (Kellgren-Lawrence grade 3 or higher), with horizontal flip augmentation applied to double the dataset to 600 samples, followed by fivefold cross-validation. An extended temporal validation was conducted on a separate cohort of 50 knee OA patients. The model's accuracy was assessed by calculating the mean absolute error between predicted and actual HKA angles. Additionally, the classification of varus deformity stages was conducted to evaluate the model's ability to provide clinically relevant categorizations. Time efficiency was compared between the automated model and manual measurements performed by an experienced orthopedic surgeon. Results The ResNet-50 model achieved a bias of − 0.025° with a standard deviation of 1.422° in the retrospective cohort and a bias of − 0.008° with a standard deviation of 1.677° in the temporal validation cohort. Using the ResNet-152 model, it accurately classified the four stages of varus deformity with weighted F1-score of 0.878 and 0.859 in the retrospective and temporal validation cohorts, respectively. The automated model was 126.7 times faster than manual measurements, reducing the total time from 49.8 min to 23.6 sec for the temporal validation cohort. Conclusions The proposed ResNet-based model provides an efficient and accurate method for measuring HKA angles and classifying varus deformity stages without the need for extensive landmark annotations. Its high accuracy and significant improvement in time efficiency make it a valuable tool for clinical practice, potentially enhancing decision-making and workflow efficiency in the management of knee OA.

Differential diagnosis of acute loss of consciousness (LOC) is crucial due to the need for different therapeutic strategies despite similar clinical presentations among etiologies such as nonconvulsive status epilepticus, metabolic encephalopathy, and benzodiazepine intoxication. While altered functional connectivity (FC) plays a pivotal role in the pathophysiology of LOC, there has been a lack of efforts to develop differential diagnosis artificial intelligence (AI) models that feature the distinctive FC change patterns specific to each LOC cause. Three approaches were applied for extracting features for the AI models: three-dimensional FC adjacency matrices, vectorized FC values, and graph theoretical measurements. Deep learning using convolutional neural networks (CNN) and various machine learning algorithms were implemented to compare classification accuracy using electroencephalography (EEG) data with different epoch sizes. The CNN model using FC adjacency matrices achieved the highest accuracy with an AUC of 0.905, with 20-s epoch data being optimal for classifying the different LOC causes. The high accuracy of the CNN model was maintained in a prospective cohort. Key distinguishing features among the LOC causes were found in the delta and theta brain wave bands. This research advances the understanding of LOC's underlying mechanisms and shows promise for enhancing diagnosis and treatment selection. Moreover, the AI models can provide accurate LOC differentiation with a relatively small amount of EEG data in 20-s epochs, which may be clinically useful.

The increased usage of IoT networks brings about new privacy risks, especially when intrusion detection systems (IDSs) rely on large datasets for machine learning (ML) tasks and depend on third parties for storing and training the ML-based IDS. This study proposes a privacy-preserving synthetic data generation method using a conditional tabular generative adversarial network (CTGAN) aimed at maintaining the utility of IoT sensor network data for IDS while safeguarding privacy. We integrate differential privacy (DP) with CTGAN by employing controlled noise injection to mitigate privacy risks. The technique involves dynamic distribution adjustment and quantile matching to balance the utility–privacy tradeoff. The results indicate a significant improvement in data utility compared to the standard DP method, achieving a KS test score of 0.80 while minimizing privacy risks such as singling out, linkability, and inference attacks. This approach ensures that synthetic datasets can support intrusion detection without exposing sensitive information.

This study develops a vision-based technique for enhancing taillight recognition in autonomous vehicles, aimed at improving real-time decision making by analyzing the driving behaviors of vehicles ahead. The approach utilizes a convolutional 3D neural network (C3D) with feature simplification to classify taillight images into eight distinct states, adapting to various environmental conditions. The problem addressed is the variability in environmental conditions that affect the performance of vision-based systems. Our objective is to improve the accuracy and generalizability of taillight signal recognition under different conditions. The methodology involves using a C3D model to analyze video sequences, capturing both spatial and temporal features. Experimental results demonstrate a significant improvement in the model′s accuracy (85.19%) and generalizability, enabling precise interpretation of preceding vehicle maneuvers. The proposed technique effectively enhances autonomous vehicle navigation and safety by ensuring reliable taillight state recognition, with potential for further improvements under nighttime and adverse weather conditions. Additionally, the system reduces latency in signal processing, ensuring faster and more reliable decision making directly on the edge devices installed within the vehicles.

The diaphragm plays a crucial role in respiration, and diaphragm dysfunction is common in COPD, contributing to worsening symptoms and higher mortality rates. Many methods have been implemented to evaluate the function of the diaphragm, but each comes with its unique set of limitations. To overcome this challenge, a novel approach was introduced to assess diaphragm function in patients with chronic obstructive pulmonary disease (COPD) using thoracic computed tomography scans. This novel approach involves generating a simulation of the diaphragm from lung DICOM slice images and calculating the dynamic change in respiratory motion by quantifying the difference in the simulated diaphragm mesh area during inhalation and exhalation. The experimental design incorporates various image processing algorithms, computational geometry algorithms, and a surface fitting algorithm optimized to minimize the potential sources of error. When the proposed technique was applied to detect diaphragmatic dysfunction in patients with COPD, the results of the Pearson correlation analysis showed a strong relationship between the variables ratio of exhalation to inhalation surface and average z-value difference inhalation and exhalation surfaces, with a coefficient of 0.731. The results suggest that the proposed technique is highly accurate and beneficial in scenarios where diaphragm function is essential, such as respiratory disorders, neuromuscular diseases like Amyotrophic Lateral Sclerosis or muscular dystrophy, spinal cord injuries, and conditions such as sleep apnea. By providing specific numerical data on sensitivity and specificity, this approach offers a quantitative evaluation of its effectiveness in detecting diaphragmatic dysfunction in patients with COPD. In conclusion, the proposed approach is currently semi-automated, however, future study may investigate fully automated approaches using this technique.

Networks within the Internet of Things (IoT) have some of the most targeted devices due to their lightweight design and the sensitive data exchanged through smart city networks. One way to protect a system from an attack is to use machine learning (ML)-based intrusion detection systems (IDSs), significantly improving classification tasks. Training ML algorithms require a large network traffic dataset; however, large storage and months of recording are required to capture the attacks, which is costly for IoT environments. This study proposes an ML pipeline using the conditional tabular generative adversarial network (CTGAN) model to generate a synthetic dataset. Then, the synthetic dataset was evaluated using several types of statistical and ML metrics. Using a decision tree, the accuracy of the generated dataset reached 0.99, and its lower complexity reached 0.05 s training and 0.004 s test times. The results show that synthetic data accurately reflect real data and are less complex, making them suitable for IoT environments and smart city applications. Thus, the generated synthetic dataset can further train models to secure IoT networks and applications.

In the era of the fourth Industrial Revolution, software has recently been applied in many fields. As the size and complexity of software increase, security attack problems continue to arise owing to potential software defects, resulting in significant social losses. To reduce software defects, a secure software development life cycle (SDLC) should be systematically developed and managed. In particular, a software weakness analyzer that uses a static analysis tool to check software weaknesses at the time of development is a very effective tool for solving software weaknesses. However, because numerous false alarms can be reported even when they are not real weaknesses, programmers and reviewers must review them, resulting in a decrease in the productivity of development. In this study, we present a system that uses the BERT model to determine the reliability of the weakness analysis results generated by the static analysis tool and to reduce false alarms by reclassifying the derived results into a decision tree model. Thus, it is possible to maintain the advantages of static analysis tools and increase productivity by reducing the cost of program development and the review process.

Today, the Internet of Things (IoT) is used for convenience in everyday life in many areas. Owing to the fact that the data collected from the IoT are generated in large quantities, cloud computing is inevitably used to store and analyze the data. However, cloud storage is not owned by the user, so it is unreliable. Verifying the integrity of data collected in an IoT environment and stored in a cloud has two problems: a large amount of data needs to be verified, and the data verification should be done directly on the IoT device. Many methods for data integrity verification use trusted third parties and devices that provide sufficient resources. However, it is difficult to directly apply existing research to the IoT devices that have limited resources. This paper proposes a secure cloud storage service for an IoT environment that is based on a provable data possession model and uses Bloom filters. The experimental results showed that the proposed method saves time and has no significant differences in the verification rate with existing methods, even though the Bloom filter causes false positives. Therefore, the proposed service can effectively process a large amount of data generated in an IoT environment.

Digital evidence, such as evidence from CCTV and event data recorders, is highly valuable in criminal investigations, and is used as definitive evidence in trials. However, there are risks when digital evidence obtained during the investigation of a case is managed through a physical hard disk drive until it is submitted to the court. Previous studies have focused on the integrated management of digital evidence in a centralized system, but if a centralized system server is attacked, major operations and investigation information may be leaked. Therefore, there is a need to reliably manage digital evidence and investigation information using blockchain technology in a distributed system environment. However, when large amounts of data—such as evidence videos—are stored in a blockchain, the data that must be processed only within one block before being created increase, causing performance degradation. Therefore, we propose a two-level blockchain system that separates digital evidence into hot and cold blockchains. In the criminal investigation process, information that frequently changes is stored in the hot blockchain, and unchanging data such as videos are stored in the cold blockchain. To evaluate the system, we measured the storage and inquiry processing performance of digital crime evidence videos according to the different capacities in the two-level blockchain system.

The Expanded Douglas⁻Peucker (EDP) polygonal approximation algorithm and its application method for the Opposite Angle-Based Exact Cell Decomposition (OAECD) are proposed for the mobile robot path-planning problem with curvilinear obstacles. The performance of the proposed algorithm is compared with the existing Douglas⁻Peucker (DP) polygonal approximation and vertical cell decomposition algorithm. The experimental results show that the path generated by the OAECD algorithm with EDP approximation appears much more natural and efficient than the path generated by the vertical cell decomposition algorithm with DP approximation.