检索结果-内蒙古大学图书馆

Robust visual Tracking Using Structure-Preserving Sparse Learning

ieee SIGNAL processing LETTERS 2017年第5期24卷 707-711页

作者： Kim, Hyuncheol Jeon, Semi Lee, Sangkeun Paik, Joonki Chung Ang Univ Dept Image Seoul 13557 South Korea Chung Ang Univ Grad Sch Adv Imaging Sci Multimedia Comp Lab Seoul 13557 South Korea Chung Ang Univ Dept Image Engn Seoul 156756 South Korea

Even though numerous visual tracking methods have been proposed to deal with image streams, it is a still challenging problem to facilitate a tracking method to accurately distinguish the target from the background without drifting under the severe appearance variation of target caused by distortion of local structures. For preserving local structures of target template datasets, we present a novel structure-preserving sparse learning algorithm by obtaining sparse coefficients under maximum margin projection-based subspace representation and by updating the sparse codes under multiple task feature selection framework. To reinforce local structures of targets, we adopted a novel optimization process using an accelerated proximal gradient shrinkage operation and an efficient stopping criterion. Experimental results demonstrate that the proposed method outperforms existing state-of-the-art tracking methods.

关键词： Discriminative sparse learning sparse representation visual tracking

来源：评论

学校读者我要写书评

暂无评论

Interest Point Detection Based on Laplacian Energy of Local image Network 2

Interest Point Detection Based on Laplacian Energy of Local ...

引用

2nd ieee International Conference on Wireless communications, Signal processing and Networking (WiSPNET)

作者： Pournami, P. N. Govindan, V. K. Natl Inst Technol Calicut Dept Comp Sci & Engn Calicut Kerala India IIIT Kottayam Dept Comp Sci & Engn Kottayam Kerala India

ISBN: (纸本)9781509044429

A point of interest is the characteristic of an image which can be robustly detected due to its well-defined position. The points of interest should be easily computable and are invariant to transformations in the image domain. This paper presents a novel interest point detector based on a graph centrality measure defined over the Laplacian energy of the local image network. The proposed method considers the local circular neighborhood of each pixel and associates this with an undirected graph or network. Euclidean distance is employed to connect pixels to form graph edges. The Laplacian energy of the graph will drop when a vertex is removed from it. This relative reduction indicates the significance of that vertex in the network. Graph vertices having higher values of Laplacian centrality are identified as central most nodes. This measure of pixel significance indicates the presence of interest points. The effectiveness and robustness of the proposed approach in detecting local points of high entropy values are demonstrated by the experiments conducted on standard datasets. The visual assessment of the identified interest points correlates with the coverage value calculated. Moreover, the proposed method shows stable results in the presence of geometric transformations such as rotation, viewpoint changes, and zooms. These results provide substantial evidence of the proposed interest point detector's utility in image processing applications such as image registration, object recognition, and image matching, etc.

关键词： Graph centrality image Registration Interest points Laplacian energy

来源：评论

学校读者我要写书评

暂无评论

A Spectral Graph Based image Coding Method 25

A Spectral Graph Based Image Coding Method

引用

25th Signal processing and communications Applications Conference (SIU)

作者： Yagan, Ali Cart Ozgen, Melanaet Tonkin Anadolu Univ Elekt Elekt Muhendisligi Bolumu Eskisehir Turkey

ISBN: (纸本)9781509064946

In this paper, a spectral graph based image coding method is proposed. In the proposed method, only the smoothing parameter in the graph adjacency matrix, graph Fourier transform (GFT) coefficients kept after hard thresholding and column indices of eigenvectors corresponding to kept coefficients are sent to the decoder side. It is advantageous in that this coding process is performed by sending much less information as compared to other graph based methods. Our results indicate that our proposed method enables image compression without deteriorating the visual quality by using very small numbers of GFT coefficients.

关键词： Signal processing on graphs graph Fourier transform image coding

来源：评论

学校读者我要写书评

暂无评论

A Neural Network Based Low-Light image Denoising Method 3

A Neural Network Based Low-Light Image Denoising Method

引用

3rd ieee International Conference on Computer and communications (ICCC)

作者： Zhang, Dan Zhao, Lei Xu, Duanqing Lu, Dongming Zhejiang Univ Dept Comp Sci & Technol Hangzhou Zhejiang Peoples R China

ISBN: (纸本)9781509063529

Low-light image denoising has been a hotspot for its great importance in solving vision applications, and various solutions have been proposed. In recent years, neural network has shown great potential in single image denoising. However, lacking ground truth data, most existing neural network based methods only train on images synthesized by Gaussian noise or Poisson noise or mixed Poisson-Gaussian noise rather than realistic low-light images. In this paper, we introduce a novel convolutional neural network based model for low-light image denoising. In particular, we also present a method to collecting realistic low-light images and our neural network model is trained on the obtained dataset. Evaluating experiment results by the subjective visual observation as well as the objective quality measures shows that the proposed neural network based method is comparable to state-of-the-art.

关键词： component denoising low-light image denoisng neural network image processing

来源：评论

学校读者我要写书评

暂无评论

Spatial Domain Multi-Focus image Fusion 25

Spatial Domain Multi-Focus Image Fusion

引用

25th Signal processing and communications Applications Conference (SIU)

作者： Toprak, Ahmet Nusret Aslantas, Voysel Erciyes Univ Bilgisayar Muhendisligi Bolumu Kayseri Turkey

ISBN: (纸本)9781509064946

One of the main problem of the optical imaging systems is limited depth of field which prevent from obtaining an all-in-focus image of the environment. This paper proposes a novel pixel-based multi-focus image fusion method making use of a series of multi-focus images to obtain an all-in-focus image. The proposed method, firstly, generate a focus map over the edges. Then, a full focus map is obtained by propagating the focus values at edge locations to the entire image. The proposed method and the state-of-the-art methods are compared in terms of both quantitative and visual evaluation. Based on the results, the proposed method outperforms the other ones.

关键词： multi-focus image fusion defocus estimation

来源：评论

学校读者我要写书评

暂无评论

Bit Allocation with visual Attention and visual Distortion Sensitivity 25

Bit Allocation with Visual Attention and Visual Distortion S...

引用

25th Signal processing and communications Applications Conference (SIU)

作者： Pak, Mesut Bayazit, Ulug Istanbul Tech Univ Bilgisayar Muhendisligi Bolumu Istanbul Turkey

ISBN: (纸本)9781509064946

In this work, a novel bit allocation method based on visual attention and distortion sensitivity is developed for JPEG2000. Although, visual attention map for an image can be measured by using well-known saliency map methods, true visual attention map can be obtained by conducting experiments to determine fixation points and their durations. A perception model might turn these duration of fixations into visual attention levels. Besides visual attention, visual distortion sensitivity may guide the bit allocation process effectively. This is because human visual system is more sensitive to the distortion around the edges than the distortion in the complex textured areas. In this work, a novel visual distortion sensitivity method that considers all edges without using a threshold for gradient magnitude and uses local entropy of gradient orientation distribution is proposed. Thus, the visual attention and the distortion sensitivity level of each code-block determine its quantization parameters. Using bit allocation based on the visual attention map provides higher subjective evaluation score than using bit allocation based on the post compression rate-distortion optimization method or on a previously proposed method based on the saliency map. Secondly, it is shown that the use of visual distortion sensitivity allows higher objective evaluation scores to be attained.

关键词： visual perception attention human computer interaction eye fixation bit allocation visual distortion sensitivity image coding

来源：评论

学校读者我要写书评

暂无评论

CS Regularized SENSE pMRI Reconstruction via Interferometric Modulation 2

CS Regularized SENSE pMRI Reconstruction via Interferometric...

引用

2nd ieee International Conference on Wireless communications, Signal processing and Networking (WiSPNET)

作者： Islam, Sheikh Rafiul Maity, Santi P. Ray, Ajoy Kumar Neotia Inst Technol Management & Sci South 24 Parganas 743368 W Bengal India Indian Inst Engn Sci & Technol Howrah 711103 W Bengal India

ISBN: (纸本)9781509044429

Compressed Sensing (CS) or Compressive Sampling offers an improved data acquisition rate for Parallel Magnetic Resonance Imaging (pMRI) systems to achieve reduced scanning time. In pMRI, the optimization of image quality is done through appropriate selection/ placement of coils according to the anatomy of the object to be imaged. This paper proposes an efficient interferometric modulation scheme for radio frequency (RF) receiver coils of parallel MRI (pMRI) that produces magnetization field over the object. The modulated magnetization field is beneficial for improving estimation accuracy of sensitivity profiles which enhance reconstruction quality at high data acquisition rate. A CS regularized sensitivity encoding approach is used as reconstruction technique in which the required MR image is provided through an iterative optimization process from the under-sampled observed k-space data. Extensive simulation results show a significant reduction in artifacts and thereby consequent visual improvement in the reconstructed MR images are achieved even when a high rate of acceleration factor is used. Simulation results also demonstrate that the proposed method outperforms some state-of-the-art pMRI methods, both in terms of subjective and objective quality assessment for the reconstructed images.

关键词： Compressed sensing interferometer parallel imaging pMRI image reconstruction SENSE

来源：评论

学校读者我要写书评

暂无评论

DETECTOR WITH FOCUS: NORMALIZING GRADIENT IN image PYRAMID 24

DETECTOR WITH FOCUS: NORMALIZING GRADIENT IN IMAGE PYRAMID

引用

24th ieee International Conference on image processing (ICIP)

作者： Kim, Yonghyun Kang, Bong-Nam Kim, Daijin POSTECH Dept Comp Sci & Engn Pohang South Korea

ISBN: (纸本)9781509021758

An image pyramid can extend many object detection algorithms to solve detection on multiple scales. However, interpolation during the resampling process of an image pyramid causes gradient variation, which is the difference of the gradients between the original image and the scaled images. Our key insight is that the increased variance of gradients makes the classifiers have difficulty in correctly assigning categories. We prove the existence of the gradient variation by formulating the ratio of gradient expectations between an original image and scaled images, then propose a simple and novel gradient normalization method to eliminate the effect of this variation. The proposed normalization method reduce the variance in an image pyramid and allow the classifier to focus on a smaller coverage. We show the improvement in three different visual recognition problems: pedestrian detection, pose estimation, and object detection. The method is generally applicable to many vision algorithms based on an image pyramid with gradients.

关键词： normalization detection gradient

来源：评论

学校读者我要写书评

暂无评论

Bag-of-Spatial Words(BoSW)Framework for Predicting SAR image Registration in Real Time Applications 7

Bag-of-Spatial Words(BoSW)Framework for Predicting SAR Image...

引用

7th International Conference on Advances in Computing and communications (ICACC)

作者： Sirisha, B. Sandhya, B. Paidimarry, Chandra Sekhar Sastry, A. S. Chandrasekhara KL Univ Dept Elect & Commun Engn Guntur AP India MVSR Engn Coll Dept Comp Sci & Engn Hyderabad Andhra Pradesh India Osmania Univ UCE Dept Elect & Commun Engn Hyderabad Andhra Pradesh India

SAR image registration is a precursor for several remote sensing applications, which need precise spatial transformation between the real time moving image and fixed off-line image. In such applications, the processing time in finding whether the moving images can be registered with fixed image constitute an overhead. Hence we have approached the problem by trying to predict if the given SAR images can be registered or not even without registering them. The proposed image registration approach incorporates a classifier into the standard pipeline of feature based image registration. The attributes for the classifier model are derived from fusing the spatial parameters of the feature detector to the descriptor vector in bag of visual words framework. (C) 2017 The Authors. Published by Elsevier B.V.

关键词： image Registration Feature Detection Feature Description Bag of visual words Classification

来源：评论

学校读者我要写书评

暂无评论

Least Square Denoising in Spectral Domain for Hyperspectral images 7

Least Square Denoising in Spectral Domain for Hyperspectral ...

引用

7th International Conference on Advances in Computing and communications (ICACC)

作者： Jayaprakash, Chippy Jacob, Naveen Varghese Renu, R. K. Sowmya, V Soman, K. P. Amrita Univ Amrita Vishwa Vidyapeetham Amrita Sch Engn Ctr Computat Engn & Networking CEN Coimbatore Tamil Nadu India

Denoising is one of the fundamental pre-processing tasks in image processing that improves the quality of the information in the image. processing of hyperspectral images requires high computational power and time. In this paper, a denoising technique based on least square weighted regularization in the spectral domain is proposed. The proposed technique is experimented on standard hyperspectral datasets and also, the performance of the proposed least square denoising in spectral domain is compared with least square weighted regularization in the spatial domain and total variation based denoising method. The obtained results in terms of computational time, Signal-to-Noise Ratio calculations and visual interpretation depicts that the proposed technique performs comparably better than the existing methods such as least square and total variation based hyperspectral image denoising. (C) 2017 The Authors. Published by Elsevier B.V.

关键词： Pre-processing technique Least square weighted regularization Spectral domain Hyperspectral image Denoising SNR Total variation

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：