Figure 1. Example of video frame extrapolation. Top is the extrapolated result, middle is the zoomed local details and bottom is the occlusion map computed with ground truth.

No abstract

No abstract

Weld defect classification in radiographic images using unified deep neural network with multi-level features Cite this article as: Lu Yang and Hongquan Jiang, Weld defect classification in radiographic images using unified deep neural network with multi-level features, Journal of Intelligent Manufacturing

GNSS time-series prediction plays an important role in the monitoring of crustal plate movement, and dam or bridge deformation, and the maintenance of global or regional coordinate frames. Deep learning is a state-of-the-art approach for extracting high-level abstract features from big data without any prior knowledge. Moreover, long short-term memory (LSTM) networks are a form of recurrent neural networks that have significant potential for processing time series. In this study, a novel prediction framework was proposed by combining a multi-scale sliding window (MSSW) with LSTM. Specifically, MSSW was applied for data preprocessing to effectively extract the feature relationship at different scales and simultaneously mine the deep characteristics of the dataset. Then, multiple LSTM neural networks were used to predict and obtain the final result by weighting. To verify the performance of MSSW-LSTM, 1000 daily solutions of the XJSS station in the Up component were selected for prediction experiments. Compared with the traditional LSTM method, our results of three groups of controlled experiments showed that the RMSE value was reduced by 2.1%, 23.7%, and 20.1%, and MAE was decreased by 1.6%, 21.1%, and 22.2%, respectively. Our results showed that the MSSW-LSTM algorithm can achieve higher prediction accuracy and smaller error, and can be applied to GNSS time-series prediction.

No abstract

This study aims to explore the usage behaviors of smartphones for learning purposes by Library and Information Science (LIS) students of mainland China. A quantitative questionnaire was used for collecting the subjects' usage behavior, their perceptions on learning experience through using smartphones, and their needs of accessing library services using smartphones. The authors analyzed the results from two major universities in mainland China. This research discovered that LIS students in mainland China used smartphones for their daily-life matters more readily than for learning purposes. They were also interested in accessing library services with smartphones, but not many of them were already using these services. The undergraduate and postgraduate respondents had some differences in their usage preferences and behaviors. This paper suspected those library services are already available but the promotion has been inadequate. There are obviously opportunities for educators, librarians, and technology providers to get some insights to improve mobile learning (m-learning) in universities and for students to learn better with mobile technology and apps. This study provides insights into the users' needs and the application of m-learning in mainland China, where scant similar studies have been conducted before.

Conventional control laws are limited predominantly to control the rotor main passing frequency component of helicopter structural response, while the remaining frequency components still make harsh vibration. The primary objective of this paper is to develop a novel hybrid controller to control multifrequency helicopter vibrations. The architecture of the feedforward–feedback hybrid control law is first proposed based on the filtered-x least mean square algorithm. Subsequently, the feedback loop called discrete model predictive sliding mode controller is designed in detail which combines discrete sliding mode control and model predictive control. Discrete sliding mode control is employed for its robustness, while the chattering phenomenon is eliminated and then the system state is steered to reach the sliding surface precisely in an optimal manner with the assistance of model predictive control. Also, the stability, robustness, and state tracking error bound of the feedback controller are analyzed. Performance enhancement of the hybrid control law is verified by simulations based on a simplified helicopter finite element model. By comparison with the multifrequency filtered-x least mean square algorithm under various cases, the results clearly demonstrate that the proposed algorithm deals with the rotor main passing frequency component and its harmonics simultaneously with faster convergence and better stability.