检索结果-内蒙古大学图书馆

18th IEEE/CVF International conference on computer vision (ICCV)

作者： Ma, Xinhong Gao, Junyu Xu, Changsheng Chinese Acad Sci CASIA Inst Automat Natl Lab Pattern Recognit NLPR Beijing Peoples R China Univ Chinese Acad Sci UCAS Sch Artificial Intelligence Beijing Peoples R China Peng Cheng Lab Shenzhen Peoples R China

ISBN: (纸本)9781665428125

Most unsupervised domain adaptation methods rely on rich prior knowledge about the source-target label set relationship, and they cannot recognize categories beyond the source classes, which limits their applicability in practical scenarios. this paper proposes a new paradigm for unsupervised domain adaptation, termed as Active Universal Domain Adaptation (AUDA), which removes all label set assumptions and aims for not only recognizing target samples from source classes but also inferring those from target-private classes by using active learning to annotate a small budget of target data. For AUDA, it is challenging to jointly adapt the model to the target domain and select informative target samples for annotations under a large domain gap and significant semantic shift. To address the problems, we propose an Active Universal Adaptation Network (AUAN). Specifically, we first introduce Adversarial and Diverse Curriculum Learning (ADCL), which progressively aligns source and target domains to classify whether target samples are from source classes. then, we propose a Clustering Non-transferable Gradient Embedding (CNTGE) strategy, which utilizes the clues of transferability, diversity, and uncertainty to annotate target informative sample, making it possible to infer labels for target samples of target-private classes. Finally, we propose to jointly train ADCL and CNTGE with target supervision to promote domain adaptation and target-private class recognition. Extensive experiments demonstrate that the proposed AUDA model equipped with ADCL and CNTGE achieves significant results on four popular benchmarks.

关键词： Budget control

来源：评论

学校读者我要写书评

暂无评论

Coconut Internal Structure Feature Detection and Developmental Quality Inspection Based on YOLOv7

Coconut Internal Structure Feature Detection and Development...

引用

pattern recognition and Artificial Intelligence (PRAI), International conference on

作者： Mingwei Che Ziheng Fu Mengxing Huang Yu Zhang Jing Chen Chengxu Sun Chun Li School of Information and Communication Engineering Hainan University Haikou China School of Computer Science and Technology Hainan University Haikou China School of Medicine Affiliated Haikou Hospital Central South University Xiangya Haikou China Chinese Academy of Tropical Agricultural Sciences Coconut Research Institute Wenchang China

Coconut is one of the largest seeded fruits in the world. While its hard shell provides some protection for the growth and development of coconuts, it also hinders the observation of internal structural changes and quality inspection during early seedling selection. the cultivation of coconut seedlings takes about a year, and it takes several years for them to grow into mature coconut trees. therefore, early selection of coconuts is crucial. To address the issue of difficult observation of internal coconut features, we utilized CT scans for non-destructive scanning to obtain internal images of coconuts at different stages. We annotated the internal structures, such as coconut water, haustorium, germ and internal cracks, which exhibit significant changes during coconut growth and development. this led to the establishment of a dataset called Internal Structure Feature of Coconuts (ISFC), providing data resources for coconut image detection. In this paper, we propose an improved model based on YOLOv7 object detection specifically designed for the ISFC dataset, aiming to detect notable changes in internal coconut features and cracks. Experimental results demonstrate that this method accurately identifies and detects the targets, achieving an average precision of up to 97.4%. It satisfies the requirements for high-precision detection and provides a basis for observing internal coconut features and conducting quality inspections during development.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Towards Accurate Scene Text recognition with Semantic Reasoning Networks

Towards Accurate Scene Text Recognition with Semantic Reason...

引用

IEEE/CVF conference on computer vision and pattern recognition (CVPR)

作者： Yu, Deli Li, Xuan Zhang, Chengquan Liu, Tao Han, Junyu Liu, Jingtuo Ding, Errui Univ Chinese Acad Sci Sch Artificial Intelligence Beijing Peoples R China Baidu Inc Dept Comp Vis Technol Vis Beijing Peoples R China Chinese Acad Sci Inst Automat Natl Lab Pattern Recognit Beijing Peoples R China Baidu Inc Beijing Peoples R China

ISBN: (纸本)9781728171685

Scene text image contains two levels of contents: visual texture and semantic information. Although the previous scene text recognition methods have made great progress over the past few years, the research on mining semantic information to assist text recognition attracts less attention, only RNN-like structures are explored to implicitly model semantic information. However, we observe that RNN based methods have some obvious shortcomings, such as time-dependent decoding manner and one-way serial transmission of semantic context, which greatly limit the help of semantic information and the computation efficiency. To mitigate these limitations, we propose a novel end-to-end trainable framework named semantic reasoning network (SRN) for accurate scene text recognition, where a global semantic reasoning module (GSRM) is introduced to capture global semantic context through multi-way parallel transmission. the state-of-the-art results on 7 public benchmarks, including regular text, irregular text and non-Latin long text, verify the effectiveness and robustness of the proposed method. In addition, the speed of SRN has significant advantages over the RNN based methods, demonstrating its value in practical use.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Adaptive Low-Light Image Enhancement with Decomposition Denoising

Adaptive Low-Light Image Enhancement with Decomposition Deno...

引用

International conference on Robotics and Automation Engineering (ICRAE)

作者： Yin Gao Chao Yan Huixiong Zeng Qiming Li Jun Li Quanzhou Institute of Equipment Manufacturing Chinese Academy of Sciences Quanzhou China Fujian Science & Technology Innovation Laboratory for Optoeletornic Information of China Fuzhou China

Noise in low-light image enhancement seriously affects human observation and computer vision algorithms, while existing methods inevitably introduce under- or over-contrast problems. We proposed an adaptive low-light image enhancement with decomposition denoising to perform a good visual effect. Firstly, the low-rank denoising algorithm is designed to decrease the noise of an input low-light image. And then, adaptive luminance enhancement is developed by generating a nonlinear function to improve the luminance of the denoise image. Finally, a color restoration method based on the relationship between spectral bands and input image into a color image. Subjective and objective experiments show that compared to several state-of-the-art methods, the proposed method can reasonably enhance image luminance and contrast while removing heavy noise.

关键词： Histograms computer vision Smoothing methods Image color analysis Noise reduction Color Visual effects

来源：评论

学校读者我要写书评

暂无评论

Preface

Lecture Notes in Computer Science (including subseries Lectu...

引用

Lecture Notes in computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 2025年 15032 LNCS卷 v-vi页

作者： Silamu, Wushouer Zha, Hongbin Zhou, Jie Liu, Cheng-Lin Lin, Zhouchen Cheng, Ming-Ming He, Ran Ubul, Kurban Peking University Beijing China Nankai University Tianjin China Chinese Academy of Sciences Beijing China Xinjiang University Xinjiang Urumqi China Xinjiang University Urumqi China Tsinghua University Beijing China

来源：评论

学校读者我要写书评

暂无评论

Robust Deep Convolutional Neural Network inspired by the Primary Visual Cortex 22

Robust Deep Convolutional Neural Network inspired by the Pri...

引用

Proceedings of the 2022 11th International conference on Computing and pattern recognition

作者： Zhanguo Dong Ming Ke Jiarong Wang Lubin Wang Gang Wang School of Computer and Communication Lanzhou University of Technology China and Beijing Institute of Basic Medical Sciences China School of Computer and Communication Lanzhou University of Technology China Beijing Institute of Basic Medical Sciences China Beijing Institute of Basic Medical Sciences China and Chinese Institute for Brain Research Beijing China

ISBN: (纸本)9781450397056

Most of the current advanced object recognition deep convolutional neural networks (DCNNs) are vulnerable to attacks of adversarial perturbations. In comparison, the primate vision system can effectively suppress the inference of adversarial perturbations. Many studies have shown that the fusion of biological vision mechanisms and DCNNs is a promising way to improve model robustness. the primary visual cortex (V1) is a key brain region for visual information processing in the biological brain, containing various simple cell orientation selection receptive fields, which can specifically respond to low-level features. therefore, we have developed an object classification DCNN model inspired by V1 orientation selection receptive fields. the V1-inspired model introduces V1 orientation selection receptive fields into DCNN through anisotropic Gaussian kernels, which can enrich the receptive fields of DCNN. In the white-box adversarial attack experiments on CIFAR-100 and Mini-ImageNet, the adversarial robustness of our model is 21.74% and 20.01% higher than that of the baseline DCNN, respectively. Compared with the SOAT VOneNet, the adversarial robustness of our model improves by 2.88% and 8.56%, respectively. It is worth pointing out that our method will not increase the parameter quantity of the baseline model, while the extra training cost is very little.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Research on Machine vision Technology for Light Industry Production

Research on Machine Vision Technology for Light Industry Pro...

引用

computer Information Science and Application Technology (CISAT), International conference on

作者： Huihui Wang Haoran Yao Yuning Xie Songyan Liu Xuqing Li Sen Qiu Liu Long Hongyu Zhao Dalian Neusoft University of Information Dalian Liaoning China Dalian Univercity of Technology Dalian Liaoning China

ISBN: (数字)9798350375107

ISBN: (纸本)9798350375114

With the rise of artificial intelligence, the process of social informatization is accelerating, leading to numerous new industries. How to use information technology to assist the upgrading of traditional industries has become a hot topic in recent years. the liquor industry is one of the traditional production industries in urgent need of transformation. Unlike other brewed beverages, the semi-open production of chinese liquor relies more on generationally passed-down experience than technology. the traditional training model lacks scientific evaluation standards, making it difficult to ensure trainees master correct techniques and scientific exertion methods. To improve the quality and efficiency of brewing action training while maintaining high-quality liquor, and to help brewing workers exert force scientifically, reduce labor intensity, and enhance labor protection, this project designed an intelligent and user-friendly brewing action training platform. Experimental results indicate that the improved model can accurately estimate human poses and effectively complete 3D human reconstruction, improving overall average accuracy by 0.74%. this validates the effectiveness of the proposed method and realizes the three-dimensional visualization of human actions. the project has preliminarily designed a simple brewing action training platform, which has practical value for the inheritance of brewing techniques and the protection of employee health.

关键词： Training Industries Solid modeling three-dimensional displays Force Data visualization Production Feature extraction Protection Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Graph Structured Network for Image-Text Matching

Graph Structured Network for Image-Text Matching

引用

IEEE/CVF conference on computer vision and pattern recognition (CVPR)

作者： Liu, Chunxiao Mao, Zhendong Zhang, Tianzhu Xie, Hongtao Wang, Bin Zhang, Yongdong Chinese Acad Sci Inst Informat Engn Beijing Peoples R China Univ Chinese Acad Sci Sch Cyber Secur Beijing Peoples R China Univ Sci & Technol China Hefei Peoples R China Xiaomi AI Lab Beijing Peoples R China

ISBN: (纸本)9781728171685

Image-text matching has received growing interest since it bridges vision and language. the key challenge lies in how to learn correspondence between image and text. Existing works learn coarse correspondence based on object co-occurrence statistics, while failing to learn fine-grained phrase correspondence. In this paper, we present a novel Graph Structured Matching Network (GSMN) to learn fine-grained correspondence. the GSMN explicitly models object, relation and attribute as a structured phrase, which not only allows to learn correspondence of object, relation and attribute separately, but also benefits to learn fine-grained correspondence of structured phrase. this is achieved by node-level matching and structure-level matching. the node-level matching associates each node with its relevant nodes from another modality, where the node can be object, relation or attribute. the associated nodes then jointly infer fine-grained correspondence by fusing neighborhood associations at structure-level matching. Comprehensive experiments show that GSMN outperforms state-of-the-art methods on benchmarks, with relative Recall@1 improvements of nearly 7% and 2% on Flickr30K and MSCOCO, respectively. Code will be released at: https://***/CrossmodalGroup/GSMN

关键词： Graphic methods

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Techniques for Lunar Impact Crater Identification Based on CCD and DEM Data 28th

Deep Learning Techniques for Lunar Impact Crater Identificat...

引用

28th International conference on Image Processing, computer vision, and pattern recognition, IPCV 2024, and 23rd International conference on Information and Knowledge Engineering, IKE 2024, held as part of the World Congress in computer Science, computer Engineering and Applied Computing, CSCE 2024

作者： Chen, Siyi Li, Cheng Ma, Yueqi Liang, Jiarui Zhu, Jionghao Tian, Xiaolin School of Computer Science and Engineering Macau University of Science and Technology Macao China School of Science and Engineering The Chinese University of Hong Kong Guangdong Shenzhen China

ISBN: (纸本)9783031859328

In recent years, the subject of detection of Lunar impact craters based on deep learning methods has been widely studied with the aim of providing efficient and effective methods for the automatic detection of Lunar impact craters. In this paper, we used the new version in the YOLO series models, YOLO v9, to apply them to the detection of Lunar impact craters based on the CCD and DEM data provided by NASA to obtain accurate detection results. the performance results indicate that YOLOv9-c achieves a precision of 73.42% on CCD data, while YOLOv9-e records a slightly higher precision of 75.27%. On DEM data, their precision is 69.43% and 70.31% respectively. Additionally, the newly developed lightweight networks, GELAN-c and GELAN-e, were tested on the same CCD and DEM datasets. GELAN-c reached a precision of 74.84% on CCD data and 70.82% on DEM data. GELAN-e showed a precision of 75.81% on CCD data and 70.87% on DEM data. the results have proved that the new version of YOLO, YOLO v9, performed effectively and excellently in the task of the Lunar impact crater. © the Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

GoogLeNet-like Model for Pedestrian Attribute Detection in Surveillance Environment

GoogLeNet-like Model for Pedestrian Attribute Detection in S...

引用

International conference on computer and Communications (ICCC)

作者： Changhong Jing Song Cao Yanyan Shen Shuqiang Wang Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Shenzhen China School of Software University of Science and Technology of China Hefei China

ISBN: (纸本)9781665409513

Social public security has attracted worldwide attention in these years. Automatic pedestrian attribute detection is of great significance to improve monitoring efficiency. Low resolution of surveillance images and multi-angle problem of image acquisition are the main challenges of current pattern recognition researches. In order to improve the pedestrian attribute detection accuracy, we proposed a GoogLeNet-like model with inception structure which replaced the optimal local construction by a series of dense substructures. the experimental results show that the proposed method outperforms the previous works.

关键词： Image resolution Computational modeling Surveillance computer architecture pattern recognition Complexity theory Security

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：