Computer vision-based action recognition of basketball players in basketball training and competition has gradually become a research hotspot. However, owing to the complex technical action, diverse background, and limb occlusion, it remains a challenging task without effective solutions or public dataset benchmarks. In this study, we defined 32 kinds of atomic actions covering most of the complex actions for basketball players and built the dataset NPU RGB+D (a large scale dataset of basketball action recognition with RGB image data and Depth data captured in Northwestern Polytechnical University) for 12 kinds of actions of 10 professional basketball players with 2169 RGB+D videos and 75 thousand frames, including RGB frame sequences, depth maps, and skeleton coordinates. Through extracting the spatial features of the distances and angles between the joint points of basketball players, we created a new feature-enhanced skeleton-based method called LSTM-DGCN for basketball player action recognition based on the deep graph convolutional network (DGCN) and long short-term memory (LSTM) methods. Many advanced action recognition methods were evaluated on our dataset and compared with our proposed method. The experimental results show that the NPU RGB+D dataset is very competitive with the current action recognition algorithms and that our LSTM-DGCN outperforms the state-of-the-art action recognition methods in various evaluation criteria on our dataset. Our action classifications and this NPU RGB+D dataset are valuable for basketball player action recognition techniques. The feature-enhanced LSTM-DGCN has a more accurate action recognition effect, which improves the motion expression ability of the skeleton data.

Fully black-box robotic testing is urgently needed given the popularity of mobile applications. A critical constraining issue for generating graphical user interface (GUI) models is identifying isomorphic GUIs. In this study, we present a deep learning-based end-to-end trainable model to determine the similarity between GUIs and subsequently identify isomorphic GUIs. The first step is to recognize the GUI element visually, and a deep learning-based objective detection method is employed. In accordance with the recognition results, the GUI skeleton is reconstructed to exclude noises from the GUI style characters. The feature vectors of the reconstructed GUI skeleton are extracted by an autoencoder, and relative entropy is employed as a metric to measure the difference between GUIs. Comparison results demonstrate the effectiveness of the proposed method.

The identification of interface elements is the first step in mobile application automated testing and the key to smooth testing. However, existing object detection algorithms have a low accuracy rate, and some tiny elements are missed in the recognition of graphical user interface (GUI) elements. To address this limitation, this paper proposes the YOLOv5-MGC algorithm, a robot vision-based interface element recognition algorithm for mobile applications. The algorithm improves the network by using K-means++ algorithm for target anchor box generation, applying the attention mechanism, adding a microscale detection layer, and introducing the Ghost bottleneck module. The proposed approach enhances the recognition accuracy of the elements through the target anchor box and attention mechanism. Moreover, it enhances the network’s ability to detect tiny elements, which improves the shortcomings of the current target detection algorithm and is conducive to further promoting mobile application robot testing and enhancing robot testing automation. Experimental results show that the YOLOv5-MGC algorithm is superior to the YOLOv5 for object detection in the recognition of GUI elements, with the mean average precision (mAP_0.5) reaching 89.8% and the recognition precision reaching 80.8%.

As an important branch of path planning, coverage path planning (CPP) is widely used for unmanned aerial vehicles (UAVs) to cover target regions with lower energy consumption. Most current works focus on convex regions, whereas others need pre-decomposition to deal with non-convex or mixed regions. Therefore, it is necessary to pursue a concise and efficient method for the latter. This paper proposes a two-stage method named Shrink-Segment by Dynamic Programming (SSDP), which aims to cover mixed regions with limited energy. First, instead of decomposing and then planning, SSDP formulates an optimal path by shrinking the rings for mixed regions. Second, a dynamic programming (DP)-based approach is used to segment the overall path for UAVs in order to meet energy limits. Experimental results show that the proposed method achieves less path overlap and lower energy consumption compared to state-of-the-art methods.

The explosive growth and rapid version iteration of various mobile applications have brought enormous workloads to mobile application testing. Robotic testing methods can efficiently handle repetitive testing tasks, which can compensate for the accuracy of manual testing and improve the efficiency of testing work. Vision-based robotic testing identifies the types of test actions by analyzing expert test videos and generates expert imitation test cases. The mobile application expert imitation testing method uses machine learning algorithms to analyze the behavior of experts imitating test videos, generates test cases with high reliability and reusability, and drives robots to execute test cases. However, the difficulty of estimating multi-dimensional gestures in 2D images leads to complex algorithm steps, including tracking, detection, and recognition of dynamic gestures. Hence, this article focuses on the analysis and recognition of test actions in mobile application robot testing. Combined with the improved YOLOv5 algorithm and the ResNet-152 algorithm, a visual modeling method of mobile application test action based on machine vision is proposed. The precise localization of the hand is accomplished by injecting dynamic anchors, attention mechanism, and the weighted boxes fusion in the YOLOv5 algorithm. The improved algorithm recognition accuracy increased from 82.6% to 94.8%. By introducing the pyramid context awareness mechanism into the ResNet-152 algorithm, the accuracy of test action classification is improved. The accuracy of the test action classification was improved from 72.57% to 76.84%. Experiments show that this method can reduce the probability of multiple detections and missed detection of test actions, and improve the accuracy of test action recognition.

Reconfiguration is a crucial method to improve the fault-tolerant capability of the integrated modular avionics (IMA) system. The quality of the blueprints has a substantial influence on operational performance after the system reconfiguration. Hence, this letter proposes an intelligent reconfiguration algorithm, which applies reinforcement learning (RL) to the IMA system. The proposed algorithm combines the critical elements of the RL model with the established IMA reconfiguration model to generate low-cost and high-quality blueprints. The experiment results demonstrate that the proposed method can not only generate feasible blueprints rapidly but also enhance system performance by optimizing load balance and reconfiguration costs.