检索结果-内蒙古大学图书馆

Generating Cartoon Images from Face Photos with Cycle-Consistent Adversarial Networks

Computers, Materials & Continua 2021年第11期69卷 2733-2747页

作者： Tao Zhang Zhanjie Zhang Wenjing Jia Xiangjian He Jie Yang School of Artificial Intelligence and Computer Science Jiangnan UniversityWuxi214000China Key Laboratory of Artificial Intelligence Jiangsu214000China The Global Big Data Technologies Centre University of Technology SydneyUltimoNSW2007Australia The Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong UniversityShanghai201100China

The generative adversarial network(GAN)is first proposed in 2014,and this kind of network model is machine learning systems that can learn to measure a given distribution of data,one of the most important applications is style *** transfer is a class of vision and graphics problems where the goal is to learn the mapping between an input image and an output ***-GAN is a classic GAN model,which has a wide range of scenarios in style *** its unsupervised learning characteristics,the mapping is easy to be learned between an input image and an output ***,it is difficult for CYCLE-GAN to converge and generate high-quality *** order to solve this problem,spectral normalization is introduced into each convolutional kernel of the *** convolutional kernel reaches Lipschitz stability constraint with adding spectral normalization and the value of the convolutional kernel is limited to[0,1],which promotes the training process of the proposed ***,we use pretrained model(VGG16)to control the loss of image content in the position of l1 *** avoid overfitting,l1 regularization term and l2 regularization term are both used in the object loss *** terms of Frechet Inception Distance(FID)score evaluation,our proposed model achieves outstanding performance and preserves more discriminative *** results show that the proposed model converges faster and achieves better FID scores than the state of the art.

关键词： Generative adversarial network spectral normalization Lipschitz stability constraint VGG16 l1 regularization term l2 regularization term Frechet inception distance

来源：评论

学校读者我要写书评

暂无评论

An Unsupervised Person Search Method for Video Surveillance 22

An Unsupervised Person Search Method for Video Surveillance

引用

8th International Conference on Computing and Artificial Intelligence, ICCAI 2022

作者： Feng, Deying Yang, Jie Wei, Yanxia Xiao, Hairong Zhang, Laigang School of Mechanical and Automotive Engineering Liaocheng University China Key Laboratory of System Control and Information Processing Ministry of Education Shanghai China Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University Shanghai China Liaocheng University China

ISBN: (纸本)9781450396110

We propose an unsupervised person search method for video surveillance. This method considers both the spatial features of persons within each frame and the temporal relationship of the same person among different frames. Thus, the spatial features are extracted by region convolutional neural network, and the temporal relationship is organized by gate recurrent unit. The spatio-temporal features are generated by the following average pooling layer and indexed by locality sensitive hashing. A surveillance video database is constructed to evaluate the proposed method, and the experimental results demonstrate that our method improves the search accuracy by utilizing the spatio-temporal features. © 2022 ACM.

关键词： Monitoring

来源：评论

学校读者我要写书评

暂无评论

Support value based fusing images with different focuses

Support value based fusing images with different focuses

引用

International Conference on Machine Learning and Cybernetics, ICMLC 2005

作者： Zheng, Sheng Sun, Yu-Qiu Tian, Jin-Wen Liu, Jian China Three Gorges University Yichang 443002 China Key Laboratory for Image Processing and Intelligent Control Institute of Pattern Recognition and Artificial Intelligence Huazhong University of Science and Technology Wuhan 430074 China

ISBN: (纸本)078039092X

Many vision-related processing tasks, including edge detection and image segmentation, can be performed more easily when all objects in the scene are in good focus. However, in practice, this may not be always feasible as optical lenses, especially those with long focal lengths, only have a limited depth of field. One classical approach to recover an everywhere-in-focus image is to use Laplacian pyramid image fusion. First, several source images with different focuses of the same scene are taken and decomposed into the low/high-frequency components image sequences. Within these decompositions, the high-frequency components image sequences with the largest magnitude are selected at each pixel location. Finally, the fused image can be recovered from the decomposed components image sequences. In the support vector machine (SVM), the pixels with larger support values have a physical meaning in the sense that they reveal relative more importance of the data points for contributing to the SVM model. In this paper, we use Laplacian pyramid for the multiresolution decomposition, and then replace the traditional salient features by support values of the mapped least squares (LS)-SVM for fusing image. Experimental results illustrate that the proposed method outperforms the traditional approach. ©2005 IEEE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Using Linear Intersection for Node Location Computation in Wireless Sensor Networks

引用

自动化学报 2006年第6期32卷 907-914页

作者： SHI Qin-Qin HUO Hong FANG Tao LI De-Ren Institute of Image Processing and Pattern Recognition Shanghai Jiaotong University Shanghai 200240 State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing Wuhan University Wuhan Hubei 430072

Knowing the locations of nodes in wireless sensor networks (WSN) is essential for many applications. Nodes in a WSN can have multiple capabilities and exploiting one or more of the capabilities can help to solve the localization problem. In this paper, we assume that each node in a WSN has the capability of distance measurement and present a location computation technique called linear intersection for node localization. We also propose an applied localization model using linear intersection and do some concerned experiments to estimate the location computation algorithm.

关键词： wirelell sensor network localization linear intersection error

来源：评论

学校读者我要写书评

暂无评论

An Improved Random Sampling LDA for Face recognition

An Improved Random Sampling LDA for Face Recognition

引用

Congress on Image and Signal processing, CISP

作者： Yunfei Jiang Xinyu Chen Ping Guo Hanqing Lu Laboratory of Image Processing and Pattern Recognition Beijing Normal University Beijing China National Laboratory of Pattern Recognition Institute of Automation Chinese Academy and Sciences Beijing China

Linear Discriminant Analysis (LDA) is one of the most used feature extraction techniques for face recognition. However, it often suffers from the small sample size problem with high dimension setting. Random Subspace Method (RSM) is a popular combining technique to improve weak classifier. Nevertheless, it remains a problem how to construct an optimal random subspace for discriminant analysis. In this paper, we propose an improved random sampling LDA for face recognition. Firstly, AdaBoost is adopted to select Gabor feature and remove redundant information. Secondly, in the selected Gabor feature space, we combine principal component analysis and RSM approaches to construct optimal random subspaces for LDA. After that, direct LDA (D-LDA) and R-LDA is applied in each subspace, respectively. Final results are obtained by combining all the LDA classifiers using a fusion rule. Experiments with both the ORL and FERET face databases demonstrate the effectiveness of our proposed method, and it shows promising results compared with previous approaches.

关键词： Sampling methods Linear discriminant analysis Face recognition Principal component analysis pattern recognition Laboratories Feature extraction Scattering Signal processing Image sampling

来源：评论

学校读者我要写书评

暂无评论

A DSP/FPGA -based parallel architecture for real-time image processing

A DSP/FPGA -based parallel architecture for real-time image ...

引用

6th World Congress on Intelligent Control and Automation, WCICA 2006

作者： Luxin, Yan Tianxu, Zhang Sheng, Zhong Institute of Pattern Recognition and Artificial Intelligence Huazhong University of Science and Technology Wuhan 430074 China Key Laboratory for Image Processing and Intelligent Control Huazhong University of Science and Technology Wuhan 430074 China

ISBN: (纸本)1424403324

DSP/FPGA-based parallel architecture oriented to real-time Image processing applications is presented. The architecture is structured with high performance DSPs interconnected by FPGA. Within FPGA a FIFO interconnection network and the specific data communication protocol are implemented, which interconnect 3 DSPs (TMS320C6414) effectively. The measured performances in the prototype with the proposed parallel architecture, including inter-DSP data communication performance and system computing capacity, show high data transfer bandwidth (up to 400 Mbytes/s) with low latency as well as high image processing performance, which achieve a good balance for parallel Image processing. © 2006 IEEE.

关键词： Digital signal processing

来源：评论

学校读者我要写书评

暂无评论

A MIL-based interactive approach for hotspot segmentation from bone scintigraphy

A MIL-based interactive approach for hotspot segmentation fr...

引用

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

作者： Shijie Geng Jingyang Ma Xiaoguang Niu Shaoyong Jia Yu Qiao Jie Yang Institute of Image Processing and Pattern Recognition The Key Laboratory of System Control and Information Processing Ministry of Education China

ISBN: (纸本)9781479999897

Bone scintigraphy is widely used to diagnose bone diseases. Accurate hotspot segmentation is a critical task for tumor metastasis diagnosis. In this paper, we propose an interactive approach to detect and extract hotspots in thoracic region based on a new multiple instance learning (MIL) method called EM-MILBoost. We convert the segmentation problem to a multiple instance learning task by constructing positive and negative bags according to the input bounding box. In order to be robust against noisy input, we train a region-level hotspot classifier with EM-MILBoost and develop several segmentation strategies based on it. The experimental results demonstrate that our method outperforms other methods and is robust against various noisy input.

关键词： Image segmentation Noise measurement Bones Training Indexes Belts Standards

来源：评论

学校读者我要写书评

暂无评论

Cattle's fur detection based on Gaussian mixture model in complex background: Application of automatic race classification of beef cattle

Cattle's fur detection based on Gaussian mixture model in co...

引用

International Conference on Advanced Computer Science and Information System (ICACSIS)

作者： Ary Noviyanto Aniati Murni Arymurthy Laboratory of Image Processing and Pattern Recognition Faculty of Computer Science Universitas Indonesia Indonesia

Segmentation becomes a difficult task if the objects and background are not homogeneous and having overlapping characteristics. Cattle segmentation from its background is required in several typical applications, such as: the automatic cattle race classification. The cattle's fur detection which is inspired from the human skin detection is investigated in this paper for cattle and background segmentation in automatic beef cattle race classification. The Gaussian mixture model that was used in skin detection has been adopted to model Bali cow and Hybrid Ongole cow in this beef cattle race classification. The RGB color space and two texture descriptors are used as the features set. The addition of texture descriptor has increased the performance of the fur detection and automatic race classification. The GMM performs well but the noise and the complexity of the background lead to misclassification.

关键词： Image color analysis Cows Accuracy Skin Feature extraction Image segmentation Entropy

来源：评论

学校读者我要写书评

暂无评论

Combining the Contrast Information with WLD for Texture Classification

Combining the Contrast Information with WLD for Texture Clas...

引用

2012 IEEE International Conference on Computer Science and Automation Engineering（CSAE 2012）

作者： Hussain Dawood Hassan Dawood Image Processing and Pattern Recognition Laboratory Beijing Normal University Beijing 100875 China

A considerable amount of research work has been done for texture classification using local or global feature extraction methods. Inspired by Weber's Law, a simple and robust Weber Local Descriptor (WLD) is a recently developed for local feature extraction. This WLD method did not consider the contrast information. In order to improve texture classification accuracy, we propose a hybrid approach that combines the WLD with contrast information in this paper. It utilizes the histogram of two complementary features WLD and the image variance calculated with the Probability Weighted Moments. Support vector machine is used for classification. The comparison of the proposed method with state of art methods like local binary pattern and WLD is experimental investigated on two publically available dataset, named as Brodatz and KTH-TIPS2-a. Results show that our proposed method outperforms over the state of art methods for texture classification.

关键词： Texture classification Weber Local Descriptor Probability Weighted Moments Support vector machine Contrast information

来源：评论

学校读者我要写书评

暂无评论

Performance comparison analysis features extraction methods for Batik recognition

Performance comparison analysis features extraction methods ...

引用

International Conference on Advanced Computer Science and Information System (ICACSIS)

作者： Ida Nurhaida Ruli Manurung Aniati Murni Arymurthy Laboratory of Pattern Recognition and Image Processing Faculty of Computer Science University of Indonesia Indonesia

Batik, as a cultural heritage from Indonesia, has a lot of motifs based on certain patterns. This paper discusses feature extraction methods for the recognition of batik motifs in digital images. In this study, the use of several feature extraction methods have been compared in terms of their performance with several scenarios for testing level accuracy. The methods include Gray Level Co-occurrence Matrices (GLCM), Canny Edge Detection, and Gabor filters. The experimental results show that the use of GLCM features has performed the best with a classification accuracy reaching 80%.

关键词： Gabor filters Image edge detection Accuracy Noise Filter banks Testing

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：