scholarly journals Multistage Polymerization Network for Multiperson Pose Estimation

2021 ◽  
Vol 2021 ◽  
pp. 1-10
Author(s):  
Yu-Fei Bai ◽  
Hong-Bo Zhang ◽  
Qing Lei ◽  
Ji-Xiang Du
Keyword(s):  
Pose Estimation ◽  
Information Exchange ◽  
Spatial Information ◽  
Complex Problem ◽  
Feature Maps ◽  
Low Level ◽  
Discriminative Feature ◽  
Level Information ◽  
Public Datasets ◽  
High Level

Multiperson pose estimation is an important and complex problem in computer vision. It is regarded as the problem of human skeleton joint detection and solved by the joint heat map regression network in recent years. The key of achieving accurate pose estimation is to learn robust and discriminative feature maps. Although the current methods have made significant progress through interlayer fusion and intralevel fusion of feature maps, few works pay attention to the combination of the two methods. In this paper, we propose a multistage polymerization network (MPN) for multiperson pose estimation. The MPN continuously learns rich underlying spatial information by fusing features within the layers. The MPN also adds hierarchical connections between feature maps at the same resolution for interlayer fusion, so as to reuse low-level spatial information and refine high-level semantic information to obtain accurate keypoint representation. In addition, we observe a lack of connection between the output low-level information and the high-level information. To solve this problem, an effective shuffled attention mechanism (SAM) is proposed. The shuffle aims to promote the cross-channel information exchange between pyramid feature maps, while attention makes a trade-off between the low-level and high-level representations of the output features. As a result, the relationship between the space and the channel of the feature map is further enhanced. Evaluation of the proposed method is carried out on public datasets, and experimental results show that our method has better performance than current methods.

scholarly journals Building Extraction in Very High Resolution Imagery by Dense-Attention Networks

2018 ◽  
Vol 10 (11) ◽  
pp. 1768 ◽  
Author(s):  
Hui Yang ◽  
Penghai Wu ◽  
Xuedong Yao ◽  
Yanlan Wu ◽  
Biao Wang ◽  
...  
Keyword(s):  
Deep Learning ◽  
High Resolution ◽  
Building Extraction ◽  
Learning Networks ◽  
Feature Maps ◽  
Low Level ◽  
High Resolution Imagery ◽  
Very High Resolution Imagery ◽  
High Level ◽  
Very High

Building extraction from very high resolution (VHR) imagery plays an important role in urban planning, disaster management, navigation, updating geographic databases, and several other geospatial applications. Compared with the traditional building extraction approaches, deep learning networks have recently shown outstanding performance in this task by using both high-level and low-level feature maps. However, it is difficult to utilize different level features rationally with the present deep learning networks. To tackle this problem, a novel network based on DenseNets and the attention mechanism was proposed, called the dense-attention network (DAN). The DAN contains an encoder part and a decoder part which are separately composed of lightweight DenseNets and a spatial attention fusion module. The proposed encoder–decoder architecture can strengthen feature propagation and effectively bring higher-level feature information to suppress the low-level feature and noises. Experimental results based on public international society for photogrammetry and remote sensing (ISPRS) datasets with only red–green–blue (RGB) images demonstrated that the proposed DAN achieved a higher score (96.16% overall accuracy (OA), 92.56% F1 score, 90.56% mean intersection over union (MIOU), less training and response time and higher-quality value) when compared with other deep learning methods.

B-Spline based globally optimal segmentation combining low-level and high-level information

2018 ◽  
Vol 73 ◽  
pp. 144-157 ◽  
Author(s):  
Shenhai Zheng ◽  
Bin Fang ◽  
Laquan Li ◽  
Mingqi Gao ◽  
Rui Chen ◽  
...  
Keyword(s):  
B Spline ◽  
Low Level ◽  
Level Information ◽  
Optimal Segmentation ◽  
High Level

scholarly journals Peer moderation of asynchronous online discussions: An exploratory study of peer e-moderating behaviour

2016 ◽  
Author(s):  
Hajar Ghadirian
Keyword(s):  
Academic Performance ◽  
Information Exchange ◽  
Exploratory Study ◽  
Knowledge Construction ◽  
Online Discussions ◽  
Low Level ◽  
Quantitative Content ◽  
Asynchronous Online Discussions ◽  
High Level ◽  
Blended Course

<p>This study explored patterns of e-moderating behaviour students performed when they were assigned as peer moderators of asynchronous online discussions in a reciprocal manner. Eighty-four students from an undergraduate blended course were observed during a 7-week-long online discussions. Using quantitative content analysis peer moderators’ interventions were analysed based on Smet, Keer, Wever, and Valcke’s (2010) scheme. The descriptive results show information exchange and knowledge construction supports were of continuous importance. Finally, a cluster analysis identified three distinct patterns of e-moderating behaviour: low-level moderators, mid-level moderators, and high-level moderators. The clusters differed in types of e-moderating support as well as their patterns of participation. High-level moderators dominated knowledge construction support and showed high level of online participation. Mid-level moderators dominated information exchange support and exhibited a moderate level of participation. Socialisation support and low level of participation were characteristics of low-level moderators. We further examined how these approaches were related to peer moderators’ perceptions of online discussions and academic performance. The results indicate that high-level moderators scored highest on all aspects of perceptions of online discussions and outperformed peer moderators in the other clusters with regard to academic performance.</p>

scholarly journals CrossFuNet: RGB and Depth Cross-Fusion Network for Hand Pose Estimation

Sensors ◽  
2021 ◽  
Vol 21 (18) ◽  
pp. 6095
Author(s):  
Xiaojing Sun ◽  
Bin Wang ◽  
Longxiang Huang ◽  
Qian Zhang ◽  
Sulei Zhu ◽  
...  
Keyword(s):  
Pose Estimation ◽  
Depth Map ◽  
Depth Information ◽  
Feature Maps ◽  
Hand Pose Estimation ◽  
Depth Sensors ◽  
Key Points ◽  
Rgb Images ◽  
Public Datasets ◽  
Hand Pose

Despite recent successes in hand pose estimation from RGB images or depth maps, inherent challenges remain. RGB-based methods suffer from heavy self-occlusions and depth ambiguity. Depth sensors rely heavily on distance and can only be used indoors, thus there are many limitations to the practical application of depth-based methods. The aforementioned challenges have inspired us to combine the two modalities to offset the shortcomings of the other. In this paper, we propose a novel RGB and depth information fusion network to improve the accuracy of 3D hand pose estimation, which is called CrossFuNet. Specifically, the RGB image and the paired depth map are input into two different subnetworks, respectively. The feature maps are fused in the fusion module in which we propose a completely new approach to combine the information from the two modalities. Then, the common method is used to regress the 3D key-points by heatmaps. We validate our model on two public datasets and the results reveal that our model outperforms the state-of-the-art methods.

Does more accessibility lead to more disclosure? Exploring the influence of information accessibility on self-disclosure in online social networks

2019 ◽  
Vol 32 (3) ◽  
pp. 754-780 ◽  
Author(s):  
Shanshan Zhang ◽  
Ron Chi-Wai Kwok ◽  
Paul Benjamin Lowry ◽  
Zhiying Liu
Keyword(s):  
Social Capital ◽  
Social Exchange ◽  
Self Disclosure ◽  
Content Type ◽  
Low Level ◽  
Information Accessibility ◽  
Different Types ◽  
Level Information ◽  
High Level ◽  
Privacy Risks

Purpose Given the importance of online social network (OSN) media features, many studies have focused on how different types of OSNs with various media features influence users’ usage and engagement. However, a recent literature review indicates that few empirical studies have considered how different types of OSNs with different information accessibility levels influence users’ beliefs and self-disclosure. By comparing two OSN platforms (OSNs with high-level information accessibility vs OSNs with low-level information accessibility), the purpose of this paper is to address this opportunity by investigating the differential impacts of the two platforms on individuals’ psychological cognition – particularly users’ social exchange beliefs – and explaining how these beliefs translate into OSN self-disclosure. Design/methodology/approach This study used a factorial design approach in an experimental setting to examine how different levels of information accessibility (high vs low), influence the social exchange beliefs (i.e. perceived social capital bridging, perceived social capital bonding and perceived privacy risks) of OSN users and subsequently influence OSN self-disclosure. Findings The results show that users on OSNs with high-level information accessibility express significantly higher perceived social capital bridging and perceived privacy risks than users on OSNs with low-level information accessibility. However, users on OSNs with low-level information accessibility express higher social bonding beliefs than users on OSNs with high-level information accessibility, indicating that there are different effect mechanisms toward OSN self-disclosure. Originality/value The focus of this research helps unveil the complex relationships between OSN design features (e.g. information accessibility), psychological cognition (e.g. social capital bridging, social capital bonding and privacy risks) and OSN self-disclosure. First, it clarifies the relationship between information accessibility and self-disclosure by examining the mediating effect of three core social exchange beliefs. Second, it uncovers the distinct effects of high-level information-accessible OSNs and low-level information-accessible OSNs on OSN self-disclosure.

Improved YOLOv3 with duplex FPN for object detection based on deep learning

2021 ◽  
pp. 002072092098352
Author(s):  
Seokyong Shin ◽  
Hyunho Han ◽  
Sang Hun Lee
Keyword(s):  
Deep Learning ◽  
Object Detection ◽  
Autonomous Vehicles ◽  
Detection Accuracy ◽  
Small Object ◽  
Feature Maps ◽  
Low Level ◽  
Small Object Detection ◽  
High Level ◽  
Networks Structure

YOLOv3 is a deep learning-based real-time object detector and is mainly used in applications such as video surveillance and autonomous vehicles. In this paper, we proposed an improved YOLOv3 (You Only Look Once version 3) applied Duplex FPN, which enhanced large object detection by utilizing low-level feature information. The conventional YOLOv3 improved the small object detection performance by applying FPN (Feature Pyramid Networks) structure to YOLOv2. However, YOLOv3 with an FPN structure specialized in detecting small objects, so it is difficult to detect large objects. Therefore, this paper proposed an improved YOLOv3 applied Duplex FPN, which can utilize low-level location information in high-level feature maps instead of the existing FPN structure of YOLOv3. This improved the detection accuracy of large objects. Also, an extra detection layer was added to the top-level feature map to prevent failure of detection of parts of large objects. Further, dimension clusters of each detection layer were reassigned to learn quickly how to accurately detect objects. The proposed method was compared and analyzed in the PASCAL VOC dataset. The experimental results showed that the bounding box accuracy of large objects improved owing to the Duplex FPN and extra detection layer, and the proposed method succeeded in detecting large objects that the existing YOLOv3 did not.

scholarly journals Advancing the Design of Self-Explanation Prompts for Complex Problem-Solving

2020 ◽  
Vol 19 (11) ◽  
pp. 88-108
Author(s):  
Hyun Joo ◽  
Jinju Lee ◽  
Dongsik Kim
Keyword(s):  
Problem Solving ◽  
Cognitive Load ◽  
Learning Outcomes ◽  
Interaction Effect ◽  
Conceptual Knowledge ◽  
Complex Problem ◽  
The Self ◽  
Complex Problem Solving ◽  
Low Level ◽  
High Level

This research investigated the effects of focus (inference vs. inference followed by integration) and level (low vs. middle vs. high) in self-explanation prompts on both cognitive load and learning outcomes. To achieve this goal, a 2*3 experiment design was employed. A total of 199 South Korean high school students were randomly assigned to one of six conditions. The two-way MANOVA was used to analyse the effects of the self-explanation prompts on learning outcomes. Results showed that there was an interaction effect between focus and level of self-explanation prompts on delayed conceptual knowledge, suggesting that the focus of self-explanation prompts could be varied depending on their level. Second, learners who were given a high level of prompts scored higher on the immediate conceptual knowledge test than those who received a low level of prompts. A two-way ANOVA was conducted to analyse the effects of the self-explanation prompts on cognitive load and showed no significant interaction effect. However, there was a main effect in the level of the prompt that a high level of self-explanation prompts imposed a lower cognitive load compared to a low level of prompts. In sum, the design and development of self-explanation prompts should consider both focus and level, especially to improve complex problem-solving skills.

scholarly journals Automated Nuclei Segmentation of Breast Cancer Histopathology

2013 ◽  
pp. 168-172
Author(s):  
Vipin Bondre ◽  
Amoli Belsare
Keyword(s):  
Breast Cancer ◽  
Bayesian Classifier ◽  
Segmentation Algorithm ◽  
Specific Information ◽  
Structural Constraints ◽  
Cell Nuclei ◽  
Low Level ◽  
Domain Specific ◽  
Level Information ◽  
High Level

Automated detection and segmentation of cell nuclei is an essential step in breast cancer histopathology, so that there is improved accuracy, speed, level of automation and adaptability to new application. The goal of this paper is to develop efficient and accurate algorithms for detecting and segmenting cell nuclei in 2-D histological images. In this paper we will implement the utility of our nuclear segmentation algorithm in accurate extraction of nuclear features for automated grading of (a) breast cancer, and (b) distinguishing between cancerous and benign breast histology specimens. In order to address the issue the scheme integrates image information across three different scales: (1) low level information based on pixel values, (2) high-level information based on relationships between pixels for object detection, and(3)domain-specific information based on relationships between histological structures. Low-level information is utilized by a Bayesian Classifier to generate likelihood that each pixel belongs to an object of interest. High-level information is extracted in two ways: (i) by a level-set algorithm, where a contour is evolved in the likelihood scenes generated by the Bayesian classifier to identify object boundaries, and (ii) by a template matching algorithm, where shape models are used to identify glands and nuclei from the low-level likelihood scenes. Structural constraints are imposed via domain specific knowledge in order to verify whether the detected objects do indeed belong to structures of interest. The efficiency of our segmentation algorithm is evaluated by comparing breast cancer grading and benign vs. cancer discrimination accuracies with corresponding accuracies obtained via manual detection and segmentation of glands and nuclei.

scholarly journals SegCloud: a novel cloud image segmentation model using deep Convolutional Neural Network for ground-based all-sky-view camera observation

2019 ◽  
Author(s):  
Wanyi Xie ◽  
Dong Liu ◽  
Ming Yang ◽  
Shaoqing Chen ◽  
Benge Wang ◽  
...  
Keyword(s):  
Image Segmentation ◽  
Cloud Cover ◽  
Weather Forecast ◽  
Cloud Detection ◽  
Feature Maps ◽  
Deep Convolutional Neural Networks ◽  
Softmax Classifier ◽  
Discrimination Ability ◽  
Low Level ◽  
High Level

Abstract. Cloud detection and cloud properties have significant applications in weather forecast, signal attenuation analysis, and other cloud-related fields. Cloud image segmentation is the fundamental and important step to derive cloud cover. However, traditional segmentation methods rely on low-level visual features of clouds, and often fail to achieve satisfactory performance. Deep Convolutional Neural Networks (CNNs) are able to extract high-level feature information of object and have become the dominant methods in many image segmentation fields. Inspired by that, a novel deep CNN model named SegCloud is proposed and applied to accurate cloud segmentation based on ground-based observation. Architecturally, SegCloud possesses symmetric encoder-decoder structure. The encoder network combines low-level cloud features to form high-level cloud feature maps with low resolution, and the decoder network restores the obtained high-level cloud feature maps to the same resolution of input images. The softmax classifier finally achieves pixel-wise classification and outputs segmentation results. SegCloud has powerful cloud discrimination ability and can automatically segment the whole sky images obtained by a ground-based all-sky-view camera. Furthermore, a new database, which includes 400 whole sky images and manual-marked labels, is built to train and test the SegCloud model. The performance of SegCloud is validated by extensive experiments, which show that SegCloud is effective and accurate for ground-based cloud segmentation and achieves better results than traditional methods. Moreover, the accuracy and practicability of SegCloud is further proved by applying it to cloud cover estimation.

scholarly journals CFAM: Estimating 3D Hand Poses from a Single RGB Image with Attention

2020 ◽  
Vol 10 (2) ◽  
pp. 618
Author(s):  
Xianghan Wang ◽  
Jie Jiang ◽  
Yanming Guo ◽  
Lai Kang ◽  
Yingmei Wei ◽  
...  
Keyword(s):  
Computer Vision ◽  
Pose Estimation ◽  
Spatial Information ◽  
Image Features ◽  
Estimation Methods ◽  
Feature Maps ◽  
Hand Pose Estimation ◽  
Rgb Images ◽  
Rgb Image ◽  
Hand Pose

Precise 3D hand pose estimation can be used to improve the performance of human–computer interaction (HCI). Specifically, computer-vision-based hand pose estimation can make this process more natural. Most traditional computer-vision-based hand pose estimation methods use depth images as the input, which requires complicated and expensive acquisition equipment. Estimation through a single RGB image is more convenient and less expensive. Previous methods based on RGB images utilize only 2D keypoint score maps to recover 3D hand poses but ignore the hand texture features and the underlying spatial information in the RGB image, which leads to a relatively low accuracy. To address this issue, we propose a channel fusion attention mechanism that combines 2D keypoint features and RGB image features at the channel level. In particular, the proposed method replans weights by using cascading RGB images and 2D keypoint features, which enables rational planning and the utilization of various features. Moreover, our method improves the fusion performance of different types of feature maps. Multiple contrast experiments on public datasets demonstrate that the accuracy of our proposed method is comparable to the state-of-the-art accuracy.

Sign in / Sign up

Export Citation Format

Share Document