检索结果-内蒙古大学图书馆

28th International Conference on Neural Information Processing, ICONIP 2021

作者： Meng, Chen Han, Run Pang, Chen Kang, Chunmeng Lyu, Chen Lyu, Lei School of Information Science and Engineering Shandong Normal University Jinan250358 China Shandong Provincial Key Laboratory for Distributed Computer Software Novel Technology Jinan250358 China

ISBN: (纸本)9783030922375

Crowd counting has important applications in many fileds, but it is still a challenging task due to background occlusion, scale variation and uneven distribution of crowd. This paper proposes the Multi-level Progressive Aggregation Network (MPANet) to enhance the channel and spatial dependencies of feature maps and effectively integrate the multi-level features. Besides, the Aggregation Refinement (AR) module is designed to integrate low-level spatial information and high-level semantic information. The proposed AR module can effectively utilize the complementary properties between multi-level features to generate high-quality density maps. Moreover, the Multi-scale Aware (MA) module is constructed to capture rich contextual information through convolutional kernels of different sizes. Furthermore, the Semantic Attention (SA) module is designed to enhance spatial and channel response on feature maps, which can reduce false recognition of the background region. Extensive experiments on four challenging datasets demonstrate that our approach outperforms most state-of-the-art methods. © 2021, Springer Nature Switzerland AG.

关键词： computer vision

来源：评论

学校读者我要写书评

暂无评论

Analysis of vocational education dynamics evolution based on topic modeling

Analysis of vocational education dynamics evolution based on...

引用

IEEE International Symposium on Information (IT) in Medicine and Education, ITME

作者： Xinhua Wang Xiaomei Yu Yunmeng Jiang Xiangwei Zheng Wei Li School of Information Science and Engineering Shandong Normal University Jinan China Shandong Provincial Key Laboratory for Distributed Computer Software Novel Technology Jinan China Shandong Normal University Library Shandong Normal University Jinan China

ISBN: (数字)9798350356670

ISBN: (纸本)9798350356687

Clarifying the focus of vocational education dynamics is crucial for understanding the current situation and future development trends. However, comprehensive exploration on vocational education dynamics is inadequate presently. In this paper, textual information on vocational education is analyzed with Latent Dirichlet Allocation, BERTopic, and the Dynamic Topic Model. Two key issues are thoroughly investigated as follows: (1) What are the main aspects of the current development of vocational education; (2) How to learn the embeddings of vocational education text and further to promote the formulation of policy. The experiments are conducted and the experimental results show that LDA and BERTopic model achieve effective performance in clustering the texts of vocational education dynamics, and the DTM captures the topics trend of vocational education, which facilitate the government to guide the formulation and implementation of policies.

关键词： Analytical models Education Government Market research Dynamic scheduling Resource management Information technology

来源：评论

学校读者我要写书评

暂无评论

Rethinking Object Saliency Ranking: A Novel Whole-flow Processing Paradigm

arXiv

引用

arXiv 2023年

作者： Song, Mengke Li, Linfeng Wu, Dunquan Song, Wenfeng Chen, Chenglizhao China University of Petroleum China Shandong Provincial Key Laboratory for Distributed Computer Software Novel Technology China Jiangsu Key Laboratory of Image and Video Understanding for Social Safety China Qingdao University China Beijing Information Science and Technology University China

Existing salient object detection methods are capable of predicting binary maps that highlight visually salient regions. However, these methods are limited in their ability to differentiate the relative importance of multiple objects and the relationships among them, which can lead to errors and reduced accuracy in downstream tasks that depend on the relative importance of multiple objects. To conquer, this paper proposes a new paradigm for saliency ranking, which aims to completely focus on ranking salient objects by their "importance order". While previous works have shown promising performance, they still face ill-posed problems. First, the saliency ranking ground truth (GT) orders generation methods are unreasonable since determining the correct ranking order is not well-defined, resulting in false alarms. Second, training a ranking model remains challenging because most saliency ranking methods follow the multi-task paradigm, leading to conflicts and trade-offs among different tasks. Third, existing regression-based saliency ranking methods are complex for saliency ranking models due to their reliance on instance mask-based saliency ranking orders. These methods require a significant amount of data to perform accurately and can be challenging to implement effectively. To solve these problems, this paper conducts an in-depth analysis of the causes and proposes a whole-flow processing paradigm of saliency ranking task from the perspective of "GT data generation", "network structure design" and "training protocol". The proposed approach outperforms existing state-of-the-art methods on the widely-used SALICON set, as demonstrated by extensive experiments with fair and reasonable comparisons. The saliency ranking task is still in its infancy, and our proposed unified framework can serve as a fundamental strategy to guide future work. The code and data will be available at https://***/MengkeSong/Saliency-Ranking-Paradigm. Copyright © 2023, The Authors. All rig

关键词： Economic and social effects

来源：评论

学校读者我要写书评

暂无评论

PC-BNA: Parallel Convolution Binary Neural Network Accelerator

PC-BNA: Parallel Convolution Binary Neural Network Accelerat...

引用

IEEE International Conference on High Performance Computing and Communications (HPCC)

作者： Haihui Liu Yuang Zhang Xiangwei Zheng Mengxin Yu Baoxu An Xiaojie Liu Shandong Provincial Key Laboratory for Distributed Computer Software Novel Technology School of Information Science and Engineering Shandong Normal University Jinan China

Binary neural network (BNN) is widely used in speech recognition, image processing and other fields to save memory and speed up computing. However, the accuracy of the existing binarization scheme in the realistic dataset is obviously low, and the input layer uses 32-bit floating point to avoid excessive precision loss, which requires additional computing units and increases the computational burden. Therefore, it is very important to improve the input layer, save computing resources and reduce precision loss. In this paper, we propose a parallel convolution binary neural network accelerator architecture (PC-BNA). Based on our proposed BNN model, we design an efficient field programmable gate array (FPGA) based accelerator. The input of the first layer is binarized, the traditional binary convolution layer is replaced by a parallel binary convolution, and the network building blocks are improved. The experimental results show that the proposed PC-BNA has higher accuracy and better performance on CIFAR-10. The image recognition accuracy is as high as 91.4%, which is superior to the state-of-art of the BNN accelerator. Compared to the state-of-the-art BNN model, when using the same model size, the use of look-up-table (LUT) saves 9.08%, and the digital signal processor (DSP) saves 27.7%. The results suggest that it is promising for PC-BNA in future high performance mobile artificial intelligence applications.

关键词： Performance evaluation Convolution Neural networks Digital signal processors Speech recognition Logic gates Mobile handsets

来源：评论

学校读者我要写书评

暂无评论

SATMask: Spatial Attention Transform Mask for Dense Instance Segmentation 7

SATMask: Spatial Attention Transform Mask for Dense Instance...

引用

7th IEEE International Conference on Data Science in Cyberspace, DSC 2022

作者： Mao, Quanzhong Sun, Lijuan Wu, Jingchen Gao, Yutong Wu, Xu Qiu, Lirong Ministry of Education Key Laboratory of Trustworthy Distributed Computing and Service China Bupt School of Cyberspace Security Bupt School of Computer Science National Pilot Software Engineering School Beijing Jiaotong University Beijing University of Posts and Telecommunications Beijing China Beijing University of Posts and Telecommunications Library China

ISBN: (数字)9781665474801

ISBN: (纸本)9781665474801

There are often dense objects in the images processed by instance segmentation, but too dense objects will cause the problem that the objects are difficult to segment. Most of the current dense instance segmentation methods are based on dense sliding window such as TensorMask. However, the sliding window has the problems of high computation and difficult design of anchor. In order to solve the above difficulties, we propose an anchor-free and single shot dense image segmentation framework, named SATMask, which adds a Spatial Attention Transform (SAT) mask head on anchor-free one stage object detector (FCOS) to predict high quality instance mask with low complexity, and uses feature-aligned pyramid network to fuse the feature map generated by backbone to obtain rich spatial details and better semantic information. Extensive experiments on the challenging COCO and Cityscapes datasets demonstrate the effectiveness of SATMask. In particular, under the same backbone (ResNet-101), SATMask achieves 39.8% AP on COCO, surpassing the state-of-the-art instance segmentation method Mask R-CNN 1.5% AP. © 2022 IEEE.

关键词： Discrete cosine transforms

来源：评论

学校读者我要写书评

暂无评论

Teaching Activity Design Based on the Discipline key Competence of IT in Secondary Vocational School

Teaching Activity Design Based on the Discipline Key Compete...

引用

International Conference on Application of Intelligent Systems in Multi-modal Information Analytics, MMIA 2020

作者： Zhang, Penglong Zhang, Yongsheng Liu, Lining Sun, Wenqian School of Information Science and Engineering Shandong Normal University ShandongJinan250358 China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology ShandongJinan250358 China

ISBN: (纸本)9783030514303

This paper, is based on the analysis of secondary vocational students learning of information technology, that’s what we call IT. In response to a new round of curriculum reform emphasizes the key competence, we sum high school IT discipline key competence and vocational education characteristics up to the secondary vocational discipline key competence, and design related teaching activity to change the secondary vocational school IT course "computer application foundation". So as to improve secondary vocational students’ IT discipline key competence, and to provide the reference for school IT teaching material writing, content design and classroom teaching. © 2021, Springer Nature Switzerland AG.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

Named Entity Recognition of Chinese Diabetic Literature by Integrating General Domain Knowledge 4

Named Entity Recognition of Chinese Diabetic Literature by I...

引用

4th IEEE International Conference on computer and Communication Engineering Technology, CCET 2021

作者： Zhang, Duochuan Beijing University of Posts and Telecommunications School of Computer National Demonstrative Software School Beijing China Key Laboratory of Trustworthy Distributed Computing and Service Beijing University of Posts and Telecommunications Ministry of Education Beijing China

ISBN: (纸本)9781665438902

In the field of named entity recognition of diabetes literature, due to the lack of large-scale and high-quality annotation data sets, the traditional named entity recognition technology has the drawback of insufficient language feature extraction and poor generalization ability. In view of the above problems, this paper proposes an entity recognition method that integrates general domain knowledge for naming entity recognition in diabetes literature. This method uses Word2vec algorithm to generate feature level vectors based on Chinese diabetes literature. On this basis, the feature vectors are fused with the feature vectors generated by the pretraining model of BERT on the Chinese Wikipedia corpus, and uses principal component analysis to reduce the redundant feature vectors and enhance the expression ability of feature vectors in the general language scope. Then the context features are extracted by bidirectional long-term and short-term memory network, and finally the entity label sequence is generated by conditional random field decoding. The experiment proves that the model which integrates general domain knowledge is better than other traditional models in the recognition of Chinese diabetes literature, and the highest F1 value is 71.08%. © 2021 IEEE.

关键词： Domain Knowledge

来源：评论

学校读者我要写书评

暂无评论

Graph Convolutional Network with Long Time Memory for Skeleton-based Action Recognition

Graph Convolutional Network with Long Time Memory for Skelet...

引用

International Conference on computer Supported Cooperative Work in Design

作者： Yanpeng Qi Chen Pang Yiliang Liu Hong Liu Lei Lyu School of Information Science and Engineering Shandong Normal University Jinan China Shandong Provincial Academy of Education Recruitment and Examination Jinan China Shandong Provincial Key Laboratory for Distributed Computer Software Novel Technology Jinan China

Skeleton-based action recognition task has been widely studied in recent years. Currently, the most popular researches use graph convolutional network (GCN) to solve this task by modeling human joints data as spatio-temporal graph. However, a large number of long-term temporal motion relationships cannot be effectively captured by GCN. Thus, recurrent neural network (RNN) is introduced to solve this defect. In this work, we propose a model namely graph convolutional network with long time memory (GCN-LTM). Specifically, there are two task streams in our proposed model: GCN stream and RNN stream, respectively. The GCN stream aims to capture the spatial motion relationships as well as the RNN stream focuses on extracting the long-term temporal patterns. In addition, we introduce the contrastive learning strategy to better facilitate feature learning between these two streams. The multiple ablation experiments have verified the feasibility of our proposed model. Numerous experiments show that the proposed model is superior to the current state-of-the-art method under two large-scale datasets including NTU-RGBD and NTU-RGBD-120.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Unlearnable 3D Point Clouds: Class-wise Transformation Is All You Need 38

Unlearnable 3D Point Clouds: Class-wise Transformation Is Al...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Wang, Xianlong Li, Minghui Liu, Wei Zhang, Hangtao Hu, Shengshan Zhang, Yechao Zhou, Ziqi Jin, Hai National Engineering Research Center for Big Data Technology and System China Services Computing Technology and System Lab China Cluster and Grid Computing Lab China Hubei Engineering Research Center on Big Data Security China Hubei Key Laboratory of Distributed System Security China School of Cyber Science and Engineering Huazhong University of Science and Technology China School of Software Engineering Huazhong University of Science and Technology China School of Computer Science and Technology Huazhong University of Science and Technology China

Traditional unlearnable strategies have been proposed to prevent unauthorized users from training on the 2D image data. With more 3D point cloud data containing sensitivity information, unauthorized usage of this new type data has also become a serious concern. To address this, we propose the first integral unlearnable framework for 3D point clouds including two processes: (i) we propose an unlearnable data protection scheme, involving a class-wise setting established by a category-adaptive allocation strategy and multi-transformations assigned to samples;(ii) we propose a data restoration scheme that utilizes class-wise inverse matrix transformation, thus enabling authorized-only training for unlearnable data. This restoration process is a practical issue overlooked in most existing unlearnable literature, i.e., even authorized users struggle to gain knowledge from 3D unlearnable data. Both theoretical and empirical results (including 6 datasets, 16 models, and 2 tasks) demonstrate the effectiveness of our proposed unlearnable framework. Our code is available at https://***/CGCL-codes/UnlearnablePC. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

ASRNet: Adversarial Segmentation and Registration Networks for Multispectral Fundus Images

引用

computer Systems Science & Engineering 2021年第3期36卷 537-549页

作者： Yanyun Jiang Yuanjie Zheng Xiaodan Sui Wanzhen Jiao Yunlong He Weikuan Jia School of Information Science and Engineering Shandong Normal UniversityJinan250358China Key Laboratory of Intelligent Computing&Information Security in Universities of Shandong Shandong Provincial Key Laboratory for Novel Distributed Computer Software TechnologyInstitute of Biomedical SciencesShandong Normal UniversityJinan250358China Shandong Provincial Hospital Affiliated to Shandong University Jinan250021China INSA Lyon University of LyonCNRSInsermVilleurbanne69621CedexFrance

Multispectral imaging (MSI) technique is often used to capture imagesof the fundus by illuminating it with different wavelengths of light. However,these images are taken at different points in time such that eyeball movementscan cause misalignment between consecutive images. The multispectral imagesequence reveals important information in the form of retinal and choroidal bloodvessel maps, which can help ophthalmologists to analyze the morphology of theseblood vessels in detail. This in turn can lead to a high diagnostic accuracy of several diseases. In this paper, we propose a novel semi-supervised end-to-end deeplearning framework called “Adversarial Segmentation and Registration Nets”(ASRNet) for the simultaneous estimation of the blood vessel segmentation andthe registration of multispectral images via an adversarial learning process. ASRNet consists of two subnetworks: (i) A segmentation module S that fulfills theblood vessel segmentation task, and (ii) A registration module R that estimatesthe spatial correspondence of an image pair. Based on the segmention-drivenregistration network, we train the segmentation network using a semi-supervisedadversarial learning strategy. Our experimental results show that the proposedASRNet can achieve state-of-the-art accuracy in segmentation and registrationtasks performed with real MSI datasets.

关键词： Deep learning deformable image registration image segmentation multispectral imaging(MSI)

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：