Skeleton-based action recognition has advanced significantly in the past decade. Among deep learning-based action recognition methods, one of the most commonly used structures is a two-stream network. This type of network extracts high-level spatial and temporal features from skeleton coordinates and optical flows, respectively. However, other features, such as the structure of the skeleton or the relations of specific joint pairs, are sometimes ignored, even though using these features can also improve action recognition performance. To robustly learn more low-level skeleton features, this paper introduces an efficient fully convolutional network to process multiple input features. The network has multiple streams, each of which has the same encoder-decoder structure. A temporal convolutional network and a co-occurrence convolutional network encode the local and global features, and a convolutional classifier decodes high-level features to classify the action. Moreover, a novel fusion strategy is proposed to combine independent feature learning and dependent feature relating. Detailed ablation studies are performed to confirm the network's robustness to all feature inputs. If more features are combined and the number of streams increases, performance can be further improved. The proposed network is evaluated on three skeleton datasets: NTU-RGB + D, Kinetics, and UTKinect. The experimental results show its effectiveness and performance superiority over state-of-the-art methods.
Skeleton-based action recognition has attracted extensive attention recently in the computer vision community. Previous studies, especially GCN-based methods, have presented remarkable improvements for this task. However, in existing GCN-based methods, global average pooling is applied to the extracted features before the classifier. This may hurt the recognition performance since it neglects the fact that not all features are equally important in the temporal dimension. To tackle this issue, in this article, we propose a feature selection network (FSN) with actor-critic reinforcement learning. Given the extracted feature sequence, FSN learns to adaptively select the most representative features and discard the ambiguous features for action recognition. In addition, conventional graph convolution is a local operation, it cannot fully capture the non-local joint dependencies that could be vital to recognize the action. Thus, we also propose a generalized graph generation module to capture latent dependencies and further propose a generalized graph convolution network (GGCN). The GGCN and FSN are combined in a three-stream recognition framework, in which different types of information from skeleton data are further fused to improve the recognition accuracy. Extensive experiments demonstrate that the proposed FSN is a flexible and effective module that can cooperate with any existing GCN-based framework to enhance the recognition accuracy, the proposed GGCN can extract richer skeleton features for skeleton-based action recognition, and our method achieves superior performance over several public datasets, e.g. 95.7 top-1 accuracy on NTU-RGB+D, 86.7 top-1 accuracy on NTU-RGB+D 120, etc.
To enterprise human resources, knowledge is like what trophic substance is to human body cells. Trophic substance is originally from food, and through digestion, absorption, and blood circulation, trophic substance is transported to each cell. This forms a perfect trophic substance supply chain of human body cells. This paper, by imitating the mechanism of trophic substance transporting process and employing the principal of bionics and the method of analogy, advances a knowledge supply chain model of enterprise human resources on bionics, with a view to clarifying the knowledge chain in an enterprise, and in favor of enterprise knowledge management and its human resource development.
The LAMOST telescope has been operating steadily for more than 10 yr and is currently undergoing an upgrade to its fiber-optic positioning system on the telescope’s focal plane. By adding a vision detection system to accurately measure the current position of the optical fibres and carry out closed-loop feedback control, the positioning accuracy of the original optical fiber system is greatly improved. Due to the limitation of the existing optical path of the telescope, to avoid blocking the optical path, the detection system camera cannot be placed on the optical axis of the telescope, which makes the camera and the optical fiber end face at a relatively inclined position. However, the angle between the camera and an object will introduce errors to the polynomial calibration method used in the visual detection system. To address this problem, by analyzing the relationship between the camera imaging principle and the polynomial calibration method, this paper studies the influence of distortion and imaging angle on polynomial calibration results under different angle conditions and proposes an improved high-precision polynomial calibration method. First, the distortion parameters are calculated using Zhang’s method, and the image distortion is removed. Then, a mapping model between oblique plane image and nonoblique plane image is established through a homography matrix. The coordinates of all oblique plane feature points are mapped one by one to the nonoblique plane positions on the image space, and a good correction effect is obtained. Finally, the calibration images from different angles are calculated and analyzed through simulation experiments and compared with Zhang’s method. The results show that this method has a 10%–40% improvement over Zhang’s method.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.