检索结果-内蒙古大学图书馆

International Conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Huadeng Wang Jiejiang Yu Bingbing Li Xipeng Pan Zhenbing Liu Rushi Lan Xiaonan Luo Guangxi Key Laboratory of Image and Graphic Intelligent Processing Guilin China School of Computer Science and Information Security Guilin University of Electronic Technology Guilin China Department of Pathology Ganzhou Municipal Hospital Ganzhou China

Accurate and automated gland segmentation on pathological images can assist pathologists in diagnosing the malignancy of colorectal adenocarcinoma. However, due to various gland shapes, severe deformation of malignant glands, and overlapping adhesions between glands. Gland segmentation has always been very challenging. To address these problems, we propose a DEA model. This model consists of two branches: the backbone encoding and decoding network and the local semantic extraction network. The backbone encoding and decoding network extracts advanced Semantic features, uses the proposed feature decoder to restore feature space information, and then enhances the boundary features of the gland through boundary enhancement attention. The local semantic extraction network uses the pre-trained DeepLabv3+ as a Local semantic-guided encoder to realize the extraction of edge features. Experimental results on two public datasets, GlaS and CRAG, confirm that the performance of our method is better than other gland segmentation methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Memristor-Based Circuit of Sequential Associative Memory with Memory Interactions and Stimulus Similarity Effect

引用

IEEE Transactions on Cognitive and Developmental Systems 2025年

作者： Xiong, Dongdong Wang, Xiaoping Jiang, Yiming Yang, Chao Jiang, Man Lai, Jingang Zeng, Zhigang Huazhong University of Science and Technology School of Artificial Intelligence and Automation Wuhan430074 China Key Laboratory of Image Processing and Intelligent Control of Education Ministry of China Wuhan430074 China Hubei Key Laboratory of Braininspired Intelligent Systems Wuhan430074 China Huazhong University of Science and Technology Institute of Artificial Intelligence School of Artificial Intelligence and Automation Key Laboratory of Image Processing and Intelligent Control of Education Ministry of China Wuhan430074 China Guangxi University of Science and Technology School of Computer Science and Technology Liuzhou545006 China Huazhong University of Science and Technology Department of Physiology School of Basic Medicine and Tongji Medical College Wuhan430030 China

Most proposed memristor-based circuits of associative memory consider various mechanisms in only one associative memory. Few works on circuit design of sequential associative memory have been reported. In this paper, a memristor-based circuit of sequential associative memory with memory interactions and stimulus similarity effect is designed. Two associative memories, the prior associative memory (PAM) and the later associative memory (LAM), are formed successively. The PAM modifies the rate of the formation of the LAM, and the LAM, in turn, affects the strength of the PAM, which are two interactions in sequential associative memory, called transfer and retroaction. In addition, the magnitudes of transfer and retroaction are determined by the similarity of the conditioned stimuli. The above functions are realized by the ring input processing module, memory module, transfer module and retroaction module. It has been identified in circuit analysis that the proposed circuit is power efficient and has good robustness. Furthermore, the proposed circuit has promising applications for fire rescue robots. © 2016 IEEE.

关键词： Associative storage

来源：评论

学校读者我要写书评

暂无评论

Lightweight image Super-Resolution via Weighted Multi-Scale Residual Network

引用

IEEE/CAA Journal of Automatica Sinica 2021年第7期8卷 1271-1280页

作者： Long Sun Zhenbing Liu Xiyan Sun Licheng Liu Rushi Lan Xiaonan Luo Guangxi Key Laboratory of Image and Graphic intelligent processing Guilin University of Electronic TechnologyGuilin 541004China National Local Joint Engineering Research Center of Satellite Navigation and Location Service Guilin University of Electronic TechnologyGuilin 541004China the College of Electrical and Information Engineering Hunan UniversityChangsha 410082China

The tradeoff between efficiency and model size of the convolutional neural network(CNN)is an essential issue for applications of CNN-based algorithms to diverse real-world *** deep learning-based methods have achieved significant improvements in image super-resolution(SR),current CNNbased techniques mainly contain massive parameters and a high computational complexity,limiting their practical *** this paper,we present a fast and lightweight framework,named weighted multi-scale residual network(WMRN),for a better tradeoff between SR performance and computational *** the modified residual structure,depthwise separable convolutions(DS Convs)are employed to improve convolutional operations’***,several weighted multi-scale residual blocks(WMRBs)are stacked to enhance the multi-scale representation *** the reconstruction subnetwork,a group of Conv layers are introduced to filter feature maps to reconstruct the final high-quality *** experiments were conducted to evaluate the proposed model,and the comparative results with several state-of-the-art algorithms demonstrate the effectiveness of WMRN.

关键词： Convolutional neural network(CNN) lightweight framework multi-scale super-resolution

来源：评论

学校读者我要写书评

暂无评论

Fast reconstruction of multiphase microstructures based on statistical descriptors

引用

Physical Review E 2022年第5期105卷 055301-055301页

作者： DongDong Chen Zhi Xu XiaoRui Wang HongJie He ZhongZhou Du JiaoFen Nan School of Computer and Communication Engineering Zhengzhou University of Light Industry Zhengzhou 450000 China Guangxi Key Laboratory of Images and Graphics Intelligent Processing Guilin University of Electronics Technology Guilin 541004 China

In this paper, we propose a hierarchical simulated annealing of erosion method (HSAE) to improve the computational efficiency of multiphase microstructure reconstruction, whose computational efficiency can be improved by an order of magnitude. Reconstruction of the two-dimensional (2D) and three-dimensional (3D) multiphase microstructures (pore, grain, and clay) based on simulated annealing (SA) and HSAE are performed. In the reconstruction of multiphase microstructure with HSAE and SA, three independent two-point correlation functions are chosen as the morphological information descriptors. The two-point cluster function which contains significant high-order statistical information is used to verify the reconstruction results. From the analysis of 2D reconstruction, it can find that the proposed HSAE technique not only improves the quality of reconstruction, but also improves the computational efficiency. The reconstructions of our proposed method are still imperfect. This is because the used two-point correlation functions contain insufficient information. For the 3D reconstruction, the two-point correlation functions of the 3D generation are in excellent agreement with those of the original 2D image, which illustrates that our proposed method is effective for the reconstruction of 3D microstructure. The comparison of the energy vs computational time between the SA and HSAE methods shows that our presented method is an order of magnitude faster than the SA method. That is because only some of the pixels in the overall hierarchy need to be considered for sampling.

关键词： Stochastic processes Stochastic processes & statistics

来源：评论

学校读者我要写书评

暂无评论

Text Detection And Recognition Algorithm For Arbitrary Shapes In Natural Scenes

Text Detection And Recognition Algorithm For Arbitrary Shape...

引用

Frontiers of Electronics, Information and Computation Technologies (ICFEICT), International Conference on

作者： Zhenrong Deng Zongyang Liu Rui Yang Rong Huang Rushi Lan Guangxi Key Laboratory of Graphics Intelligent Processing Guilin University of Electronic Technology Guilin China Nanning Research Institute Guilin University of Electronic Technology Guilin China Guangxi Jinghang UAV Co. Ltd. Nanning China

Natural scenes are rich and varied in text presentation and complex in the background. These characteristics are likely to cause problems such as missing text and difficulties in coordinate regression during detection. At the same time, the gradient disappears due to the long text sequence in the recognition stage, thus causing poor correlation before and after the text. In order to address the issue, this research suggests an end-to-end scene text detection and identification model that emphasizes spatial image data and long-range text relevance. The model uses the ABCNet network as the underlying structure. In the detection stage, this paper uses ResNet to extract image features, feature fusion with iterative FPN, convolution using convolutional blocks with coordinate information, and finally, focus the coordinate information by coordinated attention mechanism. This paper uses a combination of Transformer and LSTM coding blocks in the recognition phase to encode the detection region. It uses Transformer to obtain textual long-sequence relations, LSTM to enhance short-range information, and a lightweight attention mechanism to decode. The experimental findings demonstrate that the suggested model outperforms more widely used approaches in terms of text localisation and recognition.

关键词：

来源：评论

学校读者我要写书评

暂无评论

M3YOLOv5: Feature enhanced YOLOv5 model for mandibular fracture detection

引用

Computers in Biology and Medicine 2024年 173卷 108291-108291页

作者： Zhou, Tao Wang, Hongwei Du, Yuhu Liu, Fengzhen Guo, Yujie Lu, Huiling School of Computer Science and Engineering North Minzu University Yinchuan750021 China School of Medical Information and Engineering Ningxia Medical University Yinchuan750004 China Key Laboratory of Image and Graphics Intelligent Processing of State Ethnic Affairs Commission North Minzu University Yinchuan750021 China

Background: It is very important to detect mandibular fracture region. However, the size of mandibular fracture region is different due to different anatomical positions, different sites and different degrees of force. It is difficult to locate and recognize fracture region accurately. Methods: To solve these problems, M3YOLOv5 model is proposed in this paper. Three feature enhancement strategies are designed, which improve the ability of model to locate and recognize mandibular fracture region. Firstly, Global-Local Feature Extraction Module (GLFEM) is designed. By effectively combining Convolutional Neural Network (CNN) and Transformer, the problem of insufficient global information extraction ability of CNN is complemented, and the positioning ability of the model to the fracture region is improved. Secondly, in order to improve the interaction ability of context information, Deep-Shallow Feature Interaction Module (DSFIM) is designed. In this module, the spatial information in the shallow feature layer is embedded to the deep feature layer by the spatial attention mechanism, and the semantic information in the deep feature layer is embedded to the shallow feature layer by the channel attention mechanism. The fracture region recognition ability of the model is improved. Finally, Multi-scale Multi receptive-field Feature Mixing Module (MMFMM) is designed. Deep separate convolution chains are used in this modal, which is composed by multiple layers of different scales and different dilation coefficients. This method provides richer receptive field for the model, and the ability to detect fracture region of different scales is improved. Results: The precision rate, mAP value, recall rate and F1 value of M3YOLOv5 model on mandibular fracture CT data set are 97.18%, 96.86%, 94.42% and 95.58% respectively. The experimental results show that there is better performance about M3YOLOv5 model than the mainstream detection models. Conclusion: The M3YOLOv5 model can effectiv

关键词： Fracture

来源：评论

学校读者我要写书评

暂无评论

Small Object Detection via a Dense Connection and Feature Enhancement Network

Small Object Detection via a Dense Connection and Feature En...

引用

International Conference on Computer and Communications (ICCC)

作者： Baohua Qiang Lirui Chen Shuiping Guo Lingzhi Liu Shihao Zhang Guangxi Key Laboratory of Image and Graphic Intelligent Processing Guilin University of Electronic Technology Guilin China The 7th Research Institute of China Electrics Technology Group Corporation Guangzhou China

When detecting small objects in complex environment, the features of small objects will become blurred or even lost as the number of network layers increases. To address this problem, we constructed a Dense Connection and Feature Extraction Network, termed DCFEN. First, we designed a dense-connected multi-scale feature enhancement framework, which can effectively extract and fuse multi-scale features. Second, we constructed a dense-connected subnet using dense connections, which enhances the propagation of features and the utilization of shallow feature information, improving the detection performance of small objects. Finally, extensive experimental results demonstrated the detection precision and superiority of our method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Chinese Sentiment Analysis Model Incorporating Knowledge Graph and Deep Learning

A Chinese Sentiment Analysis Model Incorporating Knowledge G...

引用

2023 7th International Conference on Electrical, Mechanical and Computer Engineering (ICEMCE)

作者： Shengping Zhao Jinlong Chen Xingguo Qin School of Computer Science and Information Security Guilin University of Electronic Technology Guilin China School of Computer Science and Information Security Guangxi Key Laboratory of Image and Graphic Intelligent Processing Guilin University of Electronic Technology Guilin Guangxi China

With the rapid growth of social media and e-commerce, users are generating large amounts of textual review data on these platforms. Analyzing and mining these comment data is important for governments, enterprises, and consumers to make decisions. Therefore, sentiment analysis, especially for Chinese comments, has become an area of wide interest in industry and academia. Aiming at the characteristics of Chinese online reviews, this paper applies the techniques of knowledge graph and deep learning around sentence-level sentiment analysis and designs a sentiment analysis model that integrates different information such as sentiment knowledge graph and serialized annotation in a deep learning model to improve the accuracy and effectiveness of sentiment analysis of Chinese online reviews. This is important for improving products and services, understanding user needs, and making more informed business and policy decisions.

关键词：

来源：评论

学校读者我要写书评

暂无评论

PFMRC: A Unified MRC Framework for Chinese Named Entity Recognition via Semantic Feature Fusion

PFMRC: A Unified MRC Framework for Chinese Named Entity Reco...

引用

Artificial Intelligence, Robotics, and Communication (ICAIRC), IEEE International Conference on

作者： Zhenrong Deng Jinglin Zhang Yannan Xiao Shiwei Wei Guangxi Key Laboratory of Images and Graphics Intelligent Processing Guilin University of Electronic Technology Guilin China Nanning Research Institute Guilin University of Electronic Technology Guilin China School of Computer Science and Engineering Guilin University of Aerospace Technology Guilin China

ISBN: (数字)9798331531225

ISBN: (纸本)9798331531232

The landscape of natural language processing (NLP) has been revolutionized by advancements that enable the upstream training of knowledge-rich models for downstream tasks, particularly named entity recognition (NER). Traditional NER approaches, treating it as a sequence labeling problem, often fail to deeply understand the whole text and to capture syntactic nuances of text. To address these issues, we introduce PFMRC, a novel machine reading comprehension (MRC) framework that leverages the strengths of both contextual semantics and syntactic information for enhanced NER performance. PFMRC employs a dual-pronged feature extraction strategy, utilizing RoBERTa and BiGRU to encode text features for contextual semantic information and representing them as node features of the graph structure, and a weighted embedding representation to translate syntactic dependencies into graph-structured edge features. Our model further incorporates Graph Attention Networks (GAT) to harmonize node and edge features, facilitating a nuanced understanding of text. The integration of prior knowledge and the modeling of intricate syntactic relationships through Graph Attention Networks (GAT) not only improves entity recognition accuracy but also enriches the feature representation, leading to more robust NER outcomes. Extensive experiments on three Chinese datasets demonstrate PFMRC's superior performance over state-of-the-art models.

关键词： Training Accuracy Translation Semantics Refining Named entity recognition Syntactics Feature extraction Solids Robots

来源：评论

学校读者我要写书评

暂无评论

intelligent Speech Enhancement Algorithm Model Based on Recorded Classroom Audio

Intelligent Speech Enhancement Algorithm Model Based on Reco...

引用

Cognitive Computing and Complex Data (ICCD), International Conference on the

作者： Xinyun Zhang Li Shen Li Zhou Dawei Luo Dejiangmo Xiaonan Luo Guangxi Key Laboratory of Image and Graphic Intelligent Processing Guilin University of Electronic Technology Guilin China School of Physics and Electronic Information Gannan Normal University Ganzhou China School of Mathematics and Computer Science Northwest Minzu University Lanzhou China

ISBN: (数字)9798350352894

ISBN: (纸本)9798350352900

With advancements in communication technology, there is increasing demand for high-quality voice communication and accurate speech recognition. In the complex acoustic environment of a recording classroom, noise, reverberation and other interferences can seriously affect the quality and intelligibility of the picked-up target speech signal, which in turn affects the effectiveness of the subsequent speech communication and recognition, so it is crucial to enhance the target speech signal in order to satisfy people's demand for clear and recognizable speech information. In this paper, based on the network model of Deep Complex Convolutional Recurrent Network (DCCRN), we introduce the F-T-LSTM structure to more accurately describe the correlation between the time and frequency domains of speech, and at the same time, we integrate the channel attention mechanism and the spatial attention mechanism to carry out a more comprehensive feature extraction of the speech features in the complex domain, and we skillfully concentrate the computational power in the area with the most feature-rich information, so that the quality and intelligibility of speech signals under the complex domain are significantly improved. significantly improve the overall performance of speech enhancement networks in the complex domain.

关键词： Time-frequency analysis Attention mechanisms Correlation Target recognition Computational modeling Speech recognition Speech enhancement Feature extraction Complexity theory Reverberation

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：