Graph Convolutional Networks (GCNs) are widely used for skeleton-based action recognition and achieved remarkable performance. Due to the locality of graph convolution, GCNs can only utilize short-range node dependencies but fail to model long-range node relationships. In addition, existing graph convolution based methods normally use a uniform skeleton topology for all frames, which limits the ability of feature learning. To address these issues, we present the Graph Convolution Network with Self-Attention (SelfGCN), which consists of a mixing features across self-attention and graph convolution (MFSG) module and a temporal-specific spatial self-attention (TSSA) module. The MFSG module models local and global relationships between joints by executing graph convolution and self-attention branches in parallel. Its bi-directional interactive learning strategy utilizes complementary clues in the channel dimensions and the spatial dimensions across both of these branches. The TSSA module uses self-attention to learn the spatial relationships between joints of each frame in a skeleton sequence. It also models the unique spatial features of the single frames. We conduct extensive experiments on three popular benchmark datasets, NTU RGB+D, NTU RGB+D120, and Northwestern-UCLA. The results of the experiment demonstrate that our method achieves or exceeds the record accuracies on all three benchmarks. Our project website is available at https://github.com/SunPengP/SelfGCN.

The accurate and efficient detection of defective insulators is an essential prerequisite for ensuring the safety of the power grid in the new generation of intelligent electrical system inspections. Currently, traditional object detection algorithms for detecting defective insulators in images face issues such as excessive parameter size, low accuracy, and slow detection speed. To address the aforementioned issues, this article proposes an insulator defect detection model based on the lightweight Faster R-CNN (Faster Region-based Convolutional Network) model (Faster R-CNN-tiny). First, the Faster R-CNN model’s backbone network is turned into a lightweight version of it by substituting EfficientNet for ResNet (Residual Network), greatly decreasing the model parameters while increasing its detection accuracy. The second step is to employ a feature pyramid to build feature maps with various resolutions for feature fusion, which enables the detection of objects at various scales. In addition, replacing ordinary convolutions in the network model with more efficient depth-wise separable convolutions increases detection speed while slightly reducing network detection accuracy. Transfer learning is introduced, and a training method involving freezing and unfreezing the model is employed to enhance the network’s ability to detect small target defects. The proposed model is validated using the insulator self-exploding defect dataset. The experimental results show that Faster R-CNN-tiny significantly outperforms the Faster R-CNN (ResNet) model in terms of mean average precision (mAP), frames per second (FPS), and number of parameters.

Meter reading is an important part of intelligent inspection, and the current meter reading method based on target detection has problems of low accuracy and large error. In order to improve the accuracy of automatic meter reading, this paper proposes an automatic reading method for pointer-type meters based on the YOLOv5-Meter Reading (YOLOv5-MR) model. Firstly, in order to improve the detection performance of small targets in YOLOv5 framework, a multi-scale target detection layer is added to the YOLOv5 framework, and a set of Anchors is designed based on the lightning rod dial data set; secondly, the loss function and up-sampling method are improved to enhance the model training convergence speed and obtain the optimal up-sampling parameters; Finally, a new external circle fitting method of the dial is proposed, and the dial reading is calculated by the center angle algorithm. The experimental results on the self-built dataset show that the Mean Average Precision (mAP) of the YOLOv5-MR target detection model reaches 79%, which is 3% better than the YOLOv5 model, and outperforms other advanced pointer-type meter reading models.

Image retrieval is becoming more and more important due to the rapid increase of the number of images on the web. To improve the efficiency of computing the similarity of images, hashing has moved into the focus of research. This paper proposes a Deep Attention-based Hash (DAH) retrieval model, which combines an attention module and a convolutional neural network to obtain hash codes with strong representability. Our DAH has the following features: The Hamming distance between the hash codes generated by similar images is small and the Hamming distance of hash codes of dissimilar images has a larger constant value. The quantitative loss from Euclidean distance to Hamming distance is minimized. DAH has a high image retrieval precision: We thoroughly compare it with ten state-of-the-art approaches on the CIFAR-10 dataset. The results show that the Mean Average Precision (MAP) of DAH reaches more than 92% in terms of 12, 24, 36 and 48 bit hash codes on CIFAR-10, which is better than what the state-of-art methods used for comparison can deliver.

Adding an attention module to the deep convolution semantic segmentation network has significantly enhanced the network performance. However, the existing channel attention module focusing on the channel dimension neglects the spatial relationship, causing location noise to transmit to the decoder. In addition, the spatial attention module exemplified by self-attention has a high training cost and challenges in execution efficiency, making it unsuitable to handle large-scale remote sensing data. We propose an efficient vector pooling attention (VPA) module for building the channel and spatial location relationship. The module can locate spatial information better by performing a unique vector average pooling in the vertical and horizontal dimensions of the feature maps. Furthermore, it can also learn the weights directly by using the adaptive local cross-channel interaction. Multiple weight learning ablation studies and comparison experiments with the classical attention modules were conducted by connecting the VPA module to a modified DeepLabV3 network using ResNet50 as the encoder. The results show that the mIoU of our network with the addition of an adaptive local cross-channel interaction VPA module increases by 3% compared to the standard network on the MO-CSSSD. The VPA-based semantic segmentation network can significantly improve precision efficiency compared with other conventional attention networks. Furthermore, the results on the WHU Building dataset present an improvement in IoU and F1-score by 1.69% and 0.97%, respectively. Our network raises the mIoU by 1.24% on the ISPRS Vaihingen dataset. The VPA module can also significantly improve the network’s performance on small target segmentation.