Human activity recognition (HAR) is a hot research topic which aims to understand human behavior and can be applied in various applications. However, transitions between activities are usually disregarded due to their low incidence and short duration when compared against other activities, while in fact, transitions can affect the performance of the recognition system if not dealt with properly. In this paper, we propose and implement a systematic human activity recognition method to recognize basic activities (BA) and transitional activities (TA) in a continuous sensor data stream. First, raw sensor data are segmented into fragments with sliding window and the features are constructed based on window segmentation. Then, cluster analysis with K-Means is used to aggregate activity fragments into periods. Next, generally, realize the classification of BA and TA according to the shortest duration of the BA, and then deal with the hidden phenomenon of BA. Third, the fragments between adjacent BA are evaluated to decide whether they are TA or disturbance process. Finally, random forest classifier is used to accurately recognize BA and TA. The proposed method is evaluated on the public dataset SBHAR. The results demonstrate that our method effectively recognizes different activities and can deliver high accuracy with all activities considered.

Dance emotion recognition technology is of great significance for the digitalization, virtual performance, inheritance and protection of folk dance. Based on the mechanism that emotion expression in dance performance can be fully expressed through the strength and rhythm of dance movements, a novel dance emotion expression method is proposed to train hybrid deep learning neural network, to effectively identify the seven basic dance emotions of fear, anger, boredom, excitement, joy, relaxation and sadness. First, in order to fully express the emotions contained in the dance movements, this paper defines a dance emotion expression method through Laban Movement Analysis (LMA) method, which includes the characteristic parameters of the three aspects of body structure, spatial orientation and force effect, and converts the original dance movement data into three characteristic expression parameters to obtain dance emotion data. Then, Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) hybrid neural network models are used to test and train dance emotion data. Finally, in order to verify the applicability of the CNN-LSTM model, decision tree, random forest, CNN and LSTM are established and compared for accuracy. The results show that it is feasible to identify dance emotion from the perspective of dance movement, and the CNN-LSTM model is of high accuracy.

With recent advents of industrial Internet of Things (IIoT), the connectivity and data collection capabilities of industrial equipment have be significantly enhanced, yet bringing new challenges for the remaining useful life (RUL) prediction. To fulfill the RUL predicting demand in multivariate time series, this work proposes an encoder-decoder model termed as dual-scale transformer model (DSFormer), built upon the Transformer architecture. First, in the encoder part, a dual-attention module is designed for the weight feature extraction from both dimensions of the sensor and time series, aiming to compensate for the diverse impacts of different sensors on the prediction. Next, a temporal convolutional network (TCN) module is introduced to capture sequence features and alleviate the loss of positional information incurred by stacking blocks. Then, the feature decomposition module is integrated into the decoder for trend feature extraction from sequences, providing the model with additional sequence information. Finally, compared to existing models, the proposed method can obtain the superior performance in terms of the root mean square error (RMSE) and Score metrics on the FD001, FD002 and FD003 subsets of the C-MAPSS dataset, with an average improvement of 3.2% and 2.5% respectively. In particular, the ablation experiment further validates the effectiveness of proposed modules in handling multivariate time series and extracting features.

As an extremely complicated task, road detection is of vital importance for the traveling comfort and driving safety. While, high-end automobiles are already equipped with road detection function, most mid-range cars can only detect and evaluate road conditions leveraging remodeled or additional hardware devices built on vehicles, thereby constraining the road quality detection. With the growing popularity of smartphones, detections based on built-in sensors emerge. Most detections on built-in sensors, nevertheless, are on the basis of Euclidean distance, thus neglecting the correlation between characteristics in road quality, i.e., the acceleration sensor and gyroscope have obvious fluctuations when the vehicle passes through the larger pothole, and there is a connection between them. In this paper, we propose a novel road detection approach based on Mahalanobis- Taguchi system (MTS), leveraging smartphones for data collection and involving the correlation between characteristics. We implemented a simple application to collect data, processed data and classified road quality conditions. The experimental test was carried out on city roads in Xi’an, Shaanxi. Experiment results reveal that the road surface conditions, including manhole cover, pothole, and speed bump, can be well differentiated with the method based on MTS. To a certain extent, the strategy of marking road conditions to the navigation map can effectively improve not only driving experience and traveling comfort, but also driving safety, thereby providing more supports for the maintenance units.

Automatic image segmentation is vital for the computer-aided determination of treatment directions, particularly in terms of labelling lesions or infected areas. However, the manual labelling of disease regions is inconsistent and a time-consuming assignment. Meanwhile, radiologists’ comments are exceedingly subjective, regularly impacted by personal clinical encounters. To address these issues, we proposed a transformer learning strategy to automatically recognize infected areas in medical images. We firstly utilize a parallel partial decoder to aggregate high-level features and then generate a global feature map. Explicit edge attention and implicit reverse attention are applied to demonstrate boundaries and enhance their expression. Additionally, to alleviate the need for extensive labeled data, we propose a segmentation network combining propagation and transformer architectures that requires only a small amount of labeled data while leveraging fundamentally unlabeled images. The attention mechanisms are integrated within convolutional networks, keeping their global structures intact. Standalone transformers connected straightforwardly and receiving image patches can also achieve impressive segmentation performance. Our network enhanced the learning ability and attained a higher quality execution. We conducted a variety of ablation studies to demonstrate the adequacy of each modelling component. Experiments conducted across various medical imaging modalities illustrate that our model beats the most popular segmentation models. The comprehensive results also show that our transformer architecture surpasses established frameworks in accuracy while better preserving the natural variations in anatomy. Both quantitatively and qualitatively, our model achieves a higher overlap with ground truth segmentations and improved boundary adhesion.

Electric vehicles(EVs) are increasing popular as clean transportation. However electric vehicle(EV) users often encounter a shortage of electricity in the driving process. Especially when vehicle charging path planning system request reaches peak hour, it will increase the system calculation pressure. To promote penetration and popularity of electric vehicles, it is of critical importance to determine location and size of charging stations more scientifically and improve the computational power of path planning. In this paper, a method is developed for charge warning and path planning of insufficient energy EVs. The method utilizes the appropriate energy consumption factor to construct the electricity early warning model, analyzes the consumption of electricity in the vehicle in real time, and promptly warns the users when the vehicle has insufficient electric energy. Meanwhile, it combines the actual traffic information and map information to construct a network topology based on time weight, queuing mechanism, and charging calculation model to obtain path network model data, charging station queuing times and charging times as parameters. EV path planning problem is solved to minimize the total times of travel using the Dijkstra algorithm with the input path network model data, charging station queuing times and charging times as parameters. At the same time, the Spark computing framework is joined innovatively to improve the computing speed and solve the problem of path planning peak period. The Spark framework is utilized to parallelize the optimized Dijkstra algorithm. The experiment shows that the electric vehicle charging warning and path planning method provides an effective way for drivers of electric vehicles to charge.