检索结果-内蒙古大学图书馆

IEEE International Conference on Multimedia and Expo (ICME)

作者： Kai Guo Xiaokang Yang Rui Zhang Songyu Yu Institute of Image Communication and Information Processing Shanghai Key Laboratory of Digital Media Processing and Transmission Shanghai Jiaotong University Shanghai China

In learning based single image super-resolution (SR) approach, the super-resolved image are usually found or combined from training database through patch matching. But because the representation ability of small patch is limited, it is difficult to guarantee that the super-resolved image is best under global view. To tackle this problem, we propose a statistical learning method for SR with both global and local constraints. Firstly, we use maximum a posteriori (MAP) estimation with learned image priors by fields of experts (FoE) model, and regularize SR globally guided by the image priors. Secondly, for each overlapped patch, the higher-order Markov random fields (MRFs) is used to model its local relationship with corresponding high-resolution candidates, then belief propagation is used to find high-resolution image. Compared with traditional patch based learning method without global constraint, our method could not only preserve the global image structure, but also restore the local details well. Experiments verify the idea of our global and local constraint SR method.

关键词： Strontium image resolution image databases Markov random fields Belief propagation image restoration Signal resolution Spatial databases image communication Information processing

来源：评论

学校读者我要写书评

暂无评论

An improving MFCC features extraction based on FastICA algorithm plus RASTA filtering

引用

Journal of Computers 2011年第7期6卷 1477-1484页

作者： Zhao, Huan Hu, Lian Peng, Xiujuan Wang, Gangjin Yu, Fei Xu, Cheng School of Information Science and Engineering Hunan University Changsha China Jiangsu Provincial Key Lab of Image Processing and Image Communications Nanjing University of Posts and Telecommunications Nanjing China

FastICA is a kind of independent component analysis (ICA), which is robust and high performance algorithm, it can strongly remove signal correlation and ensure each signal to be independence. Through perceptual test, improving that RASTA is an idea which can denoise effectively. First, we remove signal correlation through FastICA algorithm, then we use RASTA filter to filtering the ceptral coefficients. Finally, we reduce dimension of the cepstral coefficients by the variances of cepstral coefficients in different dimension and obtain our features. By the HTK3.3, the speech feature extraction which was presented in this paper show the better robust in recognition experiment. © 2011 ACADEMY PUBLISHER.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

image super-resolution based on a novel edge sharpness prior

Image super-resolution based on a novel edge sharpness prior

引用

International Conference on Pattern Recognition

作者： Qing Yan Yi Xu Xiaokang Yang Kai Chen The Institute of Image Communication and Information Processing Shanghai Key Laboratory of Digital Media Processing and Transmition Shanghai Jiaotong University Shanghai China

ISBN: (纸本)9781467322164

Edges are of significant importance in visual resolution perception. In this paper, we propose a novel image super-resolution method by enhancing the edges in the low resolution image. We first define a new edge sharpness feature: gradient profile sharpness (GPS), which considers both the absolute magnitude and the spatial scattering of edge gradient profile. Then we learn the relationship between GPSs in high resolution images and low resolution images, and we formulate a linear GPS transform to provide gradient prior for image reconstruction. Our GPS can represent edge sharpness perceptually well. And our super-resolution method can output harmonious and faithful images with better reconstruction quality.

关键词： image edge detection Global Positioning System Spatial resolution Transforms Scattering image reconstruction

来源：评论

学校读者我要写书评

暂无评论

A new no-reference stereoscopic image quality assessment based on ocular dominance theory and degree of parallax

A new no-reference stereoscopic image quality assessment bas...

引用

International Conference on Pattern Recognition

作者： Ke Gu Guangtao Zhai Xiaokang Yang Wenjun Zhang Institute of Image Communication and Information Processing Shanghai Key Laboratory of Digital Media Processing and Transmissions Shanghai Jiaotong University Shanghai China

In recent years, while stereoscopic images were becoming widely applied and the corresponding technologies were substantially developed, very few stereoscopic image quality assessment metrics were proposed, especially under the condition that there is no reference image available. This paper proposes a new no-reference stereoscopic image quality assessment algorithm based on the ocular dominance theory and degree of parallax. All of our tests using the Toyama database draw two valuable conclusions. First, the performances of stereoscopic image quality assessment methods are significantly affected by difference of image qualities between the left and right images. Second, to offset the discriminations of different degrees of parallax caused by various 3D image contents is required indeed. Experiments and comparative studies are provided to confirm the effectiveness of our proposed new stereoscopic image quality metric.

关键词： Stereo image processing image quality Measurement Indexes Transform coding image coding Visualization

来源：评论

学校读者我要写书评

暂无评论

Blurred image Registration by Combined Invariant of Legendre Moment and Harris-Laplace Detector

Blurred Image Registration by Combined Invariant of Legendre...

引用

Pacific-Rim Symposium on image and Video Technology (PSIVT)

作者： Xiubin Dai Hui Zhang Huazhong Shu Limin Luo Tianliang Liu School of Computer Science and Engineering South-East University Nanjing China Lab of Image Processing and Image Communication Nanjing University of Posts and Telecommunications Nanjing China

Since real imaging systems are imperfect, the acquired images are always distorted by blur, scale and rotation transformation. Then the registration of these degraded images has become an important task in many applications in which the moment invariants are usually efficient tools. However, the existing methods can only deal with the slightly distorted images and have the problems of information redundancy. Besides, some methods have overlapping constraint that the images to be aligned should be fully included into the reference images. In this paper, we proposed a novel method in which a new set of combined invariants based on Legendre moment holding for blur, rotation and scale degradation simultaneously were constructed as feature descriptors, and scale-invariant Harris-Laplace detector was applied to exact feature points. The experimental results show that our method can work well without overlapping constraint, especially when the distortion is great.

关键词： Detectors image registration Feature extraction Degradation Read only memory Redundancy Pattern recognition

来源：评论

学校读者我要写书评

暂无评论

A siamese network tracking algorithm based on hierarchical attention mechanism

A siamese network tracking algorithm based on hierarchical a...

引用

2020 International Symposium on Automation, Information and Computing, ISAIC 2020

作者： Zhang, Hu Hu, Dong Qiu, Yingcan Jiangsu Province’s Key Lab of Image Procession and Image Communications Nanjing210003 China Education Ministry’s Key Lab of Broadband Wireless Communication and Sensor Network Technology Nanjing210003 China Education Ministry’s Engineering Research Center of Ubiquitous Network and Heath Service Nanjing210003 China Nanjing University of Posts and Telecommunications Nanjing210003 China

A siamese network tracking algorithm based on hierarchical attention mechanism is proposed in this paper. In order to obtain more robust target tracking results, different layer features are fused effectively. In the process of extracting features, attention mechanism is used to recalibrate the feature map, and AdaBoost algorithm is used to weight the target feature map, which improves the reliability of the response map. Besides, the Inception module is also introduced which not only increases the width of the network and the adaptability of the siamese network to the scale, but also reduces the parameters and improves the speed of network training. Experimental results show that this method can effectively solve the impact of background clutter and improve the accuracy of tracking. © 2021 Institute of Physics Publishing. All rights reserved.

关键词： Target tracking

来源：评论

学校读者我要写书评

暂无评论

Learning from Noisy labels with Coarse-to-Fine Sample Credibility Modeling 17th

Learning from Noisy Labels with Coarse-to-Fine Sample Cred...

引用

17th European Conference on Computer Vision, ECCV 2022

作者： Zhang, Boshen Li, Yuxi Tu, Yuanpeng Peng, Jinlong Wang, Yabiao Wu, Cunlin Xiao, Yang Zhao, Cairong YouTu Lab Shenzhen China Tongji University Shanghai China Key Laboratory of Image Processing and Intelligent Control Ministry of Education School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China

ISBN: (纸本)9783031250620

Training deep neural network (DNN) with noisy labels is practically challenging since inaccurate labels severely degrade the generalization ability of DNN. Previous efforts tend to handle part or full data in a unified denoising flow via identifying noisy data with a coarse small-loss criterion to mitigate the interference from noisy labels, ignoring the fact that the difficulties of noisy samples are different, thus a rigid and unified data selection pipeline cannot tackle this problem well. In this paper, we first propose a coarse-to-fine robust learning method called CREMA, to handle noisy data in a divide-and-conquer manner. In coarse-level, clean and noisy sets are firstly separated in terms of credibility in a statistical sense. Since it is practically impossible to categorize all noisy samples correctly, we further process them in a fine-grained manner via modeling the credibility of each sample. Specifically, for the clean set, we deliberately design a memory-based modulation scheme to dynamically adjust the contribution of each sample in terms of its historical credibility sequence during training, thus alleviating the effect from noisy samples incorrectly grouped into the clean set. Meanwhile, for samples categorized into the noisy set, a selective label update strategy is proposed to correct noisy labels while mitigating the problem of correction error. Extensive experiments are conducted on benchmarks of different modality, including image classification (CIFAR, Clothing1M etc.) and text recognition (IMDB), with either synthetic or natural semantic noises, demonstrating the superiority and generality of CREMA. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Efficient line-based architecture for 2-D discrete wavelet transform

引用

OPTICAL ENGINEERING 2006年第3期45卷 037004-1-037004页

作者： Xiong, Cheng-Yi Tian, Jin-Wen Liu, Jian Huazhong Univ Sci & Technol Key Lab Educ Minist Image Processing Wuhan 430074 Peoples R China Huazhong Univ Sci & Technol Intelligent Control Inst Pattern Recognit & Artif Wuhan 430074 Peoples R China S Ctr Univ Natl Coll Elect Informat Engn Wuhan 430074 Peoples R China

Efficient line-based very large scale integration architectures for the 2-D discrete wavelet transform (DWT) based on a lifting scheme, using the 9/7 wavelet filters adopted in the JPEG 2000 proposal, are proposed. The embedded decimation technique based on folding and time multiplexing was exploited to optimize the architecture, which reduces the size of buffer memory required and the amount of RAM access, and hence the occupied area and power consumption of the devices. Using this technique, a single-input, single-output architecture (SISOA) and a two-input, two-output architecture (TITOA) are proposed. The presented SISOA is designed to generate one output per clock cycle;the TITOA is designed to generate two outputs per clock cycle with the same memory requirement as that for SISOA, where the four sub-band coefficients of the transformed signal are available interleaved. Because only one line of data is required at a time, a single-port memory can be used. Performance analysis and comparison results demonstrate that the proposed method is economical of hardware cost and computation time. The advantages of the design also include short output latency, simple data flow, regularity, and scalability, as well as suitability for VLSI implementation. (c) 2006 Society of Photo-Optical Instrumentation Engineers.

关键词： 2-D DWT line-based lifting scheme embedded decimation VLSI

来源：评论

学校读者我要写书评

暂无评论

Robust object tracking with bidirectional corner matching and trajectory smoothness algorithm

Robust object tracking with bidirectional corner matching an...

引用

IEEE Workshop on Multimedia Signal processing

作者： Guoshan Wu Yi Xu Xiaokang Yang Qing Yan Ke Gu Institute of Image Communication and Information Processing Shanghai Key Laboratoryof Digital Media Processing and Transmissions Shanghai JiaoTong University Shanghai China

This paper proposes a novel method for robust object tracking. The method consists of three different components: a short term tracker, an object detector, and an online object model. For the short term tracker, we use an advanced Lucas Kanade tracker with bidirectional corner matching to capture object frame by frame. Meanwhile, statistical filtering and matching algorithm combined with haar-like feature random fern play as a detector to extract all possible object candidates in the current frame. Making use of trajectory information, the online object model decides the best target match among the candidates. And the model also trains the random fern feature adaptively online to better guide consecutive tracking. We demonstrate our method is robust to track an object in a long term and under large variations of view angle and lighting conditions. Moreover, our method is efficient to re-detect the object and keep tracking even after it's out of view or recover from heavy occlusion. To achieve state-of-the-art performance, it is highlighted that our method can be extended to multiple objects tracking application. Finally, comparisons with other state-of-the-art trackers are presented to show the robustness of our tracker.

关键词： Target tracking Trajectory Detectors Robustness Feature extraction Computer vision

来源：评论

学校读者我要写书评

暂无评论

Integrating visual saliency and consistency for re-ranking image search results

Integrating visual saliency and consistency for re-ranking i...

引用

IEEE International Conference on image processing

作者： Jun Huang Xiaokang Yang Rui Zhang Fuxiang Lu Xiangzhong Fang Institute of Image Communication and Information Processing Shanghai Key Laboratory of Digital Media Processing and Transmissions Shanghai Jiaotong University Shanghai China

The paper investigates two mechanisms, visual consistency and visual saliency, in web image search: (1) In most current web image search engines, such as Google image Search and Yahoo image Search, the images that closely related to the search query are typically visually similar. These visually consistent images which occur most frequently in the first few web pages will be given higher ranks. (2) From visual aspect, it is obvious that salient images would be easier to catch users' eyes and more likely to be clicked than the cluttered ones in low-level vision. In addition, we also observe the fact that the visually salient images in the front pages are often relevant to the user's query. The principal novelty of this paper is in combining visual saliency and consistency to re-rank the results from search engines to make the re-ranked images more satisfying in both vision and content. The experimental results on a real world web image dataset demonstrate that our approach can effectively improve the performance of image retrieval.

关键词： Visualization Search engines image edge detection image color analysis image retrieval Feature extraction Computer vision

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：