Purpose: COVID-19 has become global threaten. CT acts as an important method of diagnosis. However, human–based interpretation of CT imaging is time consuming. More than that, substantial inter-observer-variation cannot be ignored. We aim at developing a diagnostic tool for artificial intelligence (AI)–based classification of CT images for recognizing COVID-19 and other common infectious diseases of the lung. Experimental Design: In this study, images were retrospectively collected and prospectively analyzed using machine learning. CT scan images of the lung that show or do not show COVID-19 were used to train and validate a classification framework based on convolutional neural network. Five conditions including COVID-19 pneumonia, non-COVID-19 viral pneumonia, bacterial pneumonia, pulmonary tuberculosis, and normal lung were evaluated. Training and validation set of images were collected from Wuhan Jin Yin-Tan Hospital whereas test set of images were collected from Zhongshan Hospital Xiamen University and the fifth Hospital of Wuhan. Results: Accuracy, sensitivity, and specificity of the AI framework were reported. For test dataset, accuracies for recognizing normal lung, COVID-19 pneumonia, non-COVID-19 viral pneumonia, bacterial pneumonia, and pulmonary tuberculosis were 99.4%, 98.8%, 98.5%, 98.3%, and 98.6%, respectively. For the test dataset, accuracy, sensitivity, specificity, PPV, and NPV of recognizing COVID-19 were 98.8%, 98.2%, 98.9%, 94.5%, and 99.7%, respectively. Conclusions: The performance of the proposed AI framework has excellent performance of recognizing COVID-19 and other common infectious diseases of the lung, which also has balanced sensitivity and specificity.

In-situ self-reporting on smartwatches has been widely-used to collect data about participants’ experience. Yet it places a burden on participants by requiring immediate response in context. Such user burden can be studied through users’ performance and perceptions. This paper evaluates six interaction techniques to study user burden of in-situ self-reporting from both performance and perception aspects under three simulated scenarios (i.e., gaming, social chatting, and walking) through an in-lab experiment with twenty-four participants. Findings show that user performance was not always aligned with their perceived burden owing to factors such as users’ acceptable interaction technique features for self-reporting, the context of use, and users’ acquired skills over time. This study also sheds lights onto how user burden is experienced across different stages of a self-reporting microinteraction. Reflecting on findings, we discuss the implications for understanding and minimizing user burden related to in-situ self-reporting and as well as other types of microinteractions.

This study addressed the insufficient structural strength and inadequate disaster resistance in building designs in the mountainous regions of Northern Vietnam. By integrating Japanese construction experience with local conditions, we proposed optimized building structures and simplified safety evaluation methods. Through an analysis of climate, terrain, geological hazards, soil conditions, and construction material costs, building design and foundation construction were optimized, and these optimizations were validated through wind tunnel experiments and finite element analysis. The results indicated that the optimized structures exhibited superior wind load stability, with external wind pressure coefficients ranging from −1.5 to −0.7, compared with the traditional structure’s range of −1 to −3.5. The redesigned foundation improved landslide resistance, reducing excavation and foundation construction costs relative to Japanese methods. The foundation’s safety factor reached 4.42–5.13, surpassing the standard of 2.5, and the retaining wall’s sliding resistance safety factor reached 1.87, exceeding the requirement of 1.5. These enhancements dramatically boosted building safety under extreme weather conditions. This study provides practical solutions for building design in Vietnam’s mountainous regions and serves as a valuable reference for similar research in other developing countries, underscoring significant practical and social implications.

Optical see-through head-mounted displays (OHMDs) blend digital content with the physical world, presenting unique color management challenges. Previous literature suggests using green as the main color, but this severely limits creative freedom. To address this, we conducted an empirical study with 30 participants, evaluating 216 colors under various OHMD usage conditions. Based on the results, we propose color guidelines indicating each hue’s clear and comfortable saturation and brightness ranges, along with clarity and comfort scores across hues for different devices and lighting conditions. Our color guidelines expand the usable color palette, offering designers a wider range of color options. These guidelines were used and iteratively refined through feedback in a workshop with 12 designers, integrating them into practical design workflows. The resulting comprehensive color guide provides a valuable resource for OHMD interface designers, enhancing both the aesthetic possibilities and functional effectiveness of augmented reality experiences.

Traditional user interface layout design heavily relies on designers' subjective judgment and lacks effective quantitative assessment methods. This study aims to overcome this shortcoming. Therefore, this paper proposes a multi-objective optimization (MOO) system designed for the low-fidelity prototyping phase to enable faster production of high-quality interface layouts. A hybrid algorithm (SNSGA-II) is designed in the system, which combines the Non-Dominated Sorting Genetic Algorithm II (NSGA-II) and the simulated annealing (SA) algorithm, along with a scoring system including both esthetics and usability. During the high-temperature phase, the SA algorithm is responsible for global exploration, while the genetic algorithm is introduced during the low-temperature phase for refined optimization, combined with an improved sigmoid function to dynamically balance the weights of each evaluation dimension. The experimental results show that this system achieved significant optimization effects across three UI design tasks with varying complexity levels. Compared with traditional NSGA-II algorithm, the SNSGA-II algorithm demonstrated approximately 50% improvement in convergence speed. Evaluation by 20 professional UI designers showed that system-optimized designs exhibited consistent improvements across all complexity levels: overall design quality enhanced by 3.0%, 6.7%, and 6.5% for simple, moderate, and complex tasks respectively; esthetic scores correspondingly increased by 5.6%, 5.7%, and 5.2%; learnability indicators improved by 4.1%, 6.1%, and 7.2% respectively. This study confirmed the effectiveness of this approach in balancing esthetics and usability, demonstrating its algorithmic performance and practical optimization potential.