检索结果-内蒙古大学图书馆

IEEE Conference on Computer Vision and Pattern Recognition

作者： Yuhua Chen Dengxin Dai Jordi Pont-Tuset Luc Van Gool Computer Vision Lab ETH Zurich

ISBN: (纸本)9781467388528

image segmentation is a key component in many computer vision systems, and it is recovering a prominent spot in the literature as methods improve and overcome their limitations. The outputs of most recent algorithms are in the form of a hierarchical segmentation, which provides segmentation at different scales in a single tree-like structure. Commonly, these hierarchical methods start from some low-level features, and are not aware of the scale information of the different regions in them. As such, one might need to work on many different levels of the hierarchy to find the objects in the scene. This work tries to modify the existing hierarchical algorithm by improving their alignment, that is, by trying to modify the depth of the regions in the tree to better couple depth and scale. To do so, we first train a regressor to predict the scale of regions using mid-level features. We then define the anchor slice as the set of regions that better balance between over-segmentation and under-segmentation. The output of our method is an improved hierarchy, re-aligned by the anchor slice. To demonstrate the power of our method, we perform comprehensive experiments, which show that our method, as a post-processing step, can significantly improve the quality of the hierarchical segmentation representations, and ease the usage of hierarchical image segmentation to high-level vision tasks such as object segmentation. We also prove that the improvement generalizes well across different algorithms and datasets, with a low computational cost.

关键词： image segmentation Computer Vision systems Hierarchical Boat anchors segmentation Object segmentation Social Hierarchy trees (mathematics)

来源：评论

学校读者我要写书评

暂无评论

Piecewise-polynomial curve fitting using group sparsity

Piecewise-polynomial curve fitting using group sparsity

引用

International Conference on Ultra Modern Telecommunications and Workshops (ICUMT)

作者： Michaela Novosadová Pavel Rajmic Dept. of Telecommunications Brno University of Technology Brno Czech Republic

ISBN: (纸本)9781467388191

We present a method for segmenting one-dimensional signal, whose segments are modeled as polynomials, and which is corrupted by an additive noise. The method is based on sparse modeling and its formulation as a convex optimization problem is solved by proximal algorithms. We perform experiments on simulated data, discuss the results and suggest future directions that could lead to even better results.

关键词： Convergence image segmentation TV Convex functions Telecommunications Signal processing Iterative methods

来源：评论

学校读者我要写书评

暂无评论

Real-time panorama composition for video surveillance using GPU

Real-time panorama composition for video surveillance using ...

引用

International Conference on Advances in Computing, Communications and Informatics (ICACCI)

作者： Pritam Prakash Shete Dinesh Madhukar Sarode Surojit Kumar Bose Computer Division BARC Mumbai India

ISBN: (纸本)9781509020300

image stitching algorithms combine multiple low resolution images and provide a single high resolution composite image with a larger field of view available for video surveillance. In this research work, we put forward and realize real-time panorama composition for a video surveillance application using the power of a GPU. We utilize a cross platform OpenGL graphics library for real-time online image processing. We parallelize panorama composition using OpenGL objects such as texture object, vertex buffer object and framebuffer object for image warping as well as edge blending to create a seamless panoramic image. We divide our panorama composition algorithm into two stages for image sources with fixed relative positions with each other. Initially in an offline stage, we compute inverse lookup maps and feather weight masks using an OpenCV image processing library for each input image. Subsequently in an online stage, we utilize these inverse lookup maps to generate warped images, which are further edge blended with each other using feather weight masks with the help of OpenGL objects. Our panorama composition is more than 8.5 times faster than the CUDA optimized OpenCV realization. It produces a high resolution seamless panoramic image using nine input image streams each with 800×600 image resolution at about more than 75 frames per second using less than 90MB of GPU memory.

关键词： Graphics processing units Feathers image resolution image edge detection Real-time systems image color analysis Streaming media

来源：评论

学校读者我要写书评

暂无评论

An optimized hardware architecture of 4×4, 8×8, 16×16 and 32×32 inverse transform for HEVC

An optimized hardware architecture of 4×4, 8×8, 16×16 and...

引用

International Conference on Advanced Technologies for Signal and image processing (ATSIP)

作者： Manel Kammoun Emna Maamouri Ahmed Ben Atitallah Nouri Masmoudi LETI laboratory—ENIS University of Sfax Sfax TUNISIA

ISBN: (纸本)9781467385275

Most video coding standards use transform algorithms to reduce the size of data characterizing a video signal. The traditional transform matrices as in H.264 are limited to 4×4 and 8×8 sizes. However, the flexibility of coding structure presented in the next generation of video coding standard High Efficiency Video Coding standard HEVC allows the definition of various sizes of transform matrices which can vary from 4×4 to 32×32. This paper describes a unified hardware architecture for 4×4, 8×8, 16×16 and 32×32 inverse 2D core transform IDCT in HEVC standard. It used only one block 1D transform and a transpose buffer based on FIFO memory blocs instead of the traditional register array in order to further reduce the memory resources. The synthesis results under TSMC 180 nm CMOS technology show that the total gate count of the design is more than 30% improved compared to previous works. However, the operating frequency of the hardware design is about 130 MHz. This last can perform the decoding of 25 frames per second of Quad HD (3840×2160) resolution.

关键词： image processing Decision support systems Hafnium

来源：评论

学校读者我要写书评

暂无评论

The Visualization of Large Graphs Accelerated by the Parallel Nearest Neighbors Algorithm

The Visualization of Large Graphs Accelerated by the Paralle...

引用

IEEE International Conference on Multimedia Big Data (BigMM)

作者： Vojtech Uher Petr Gajdo Václav Snáel Department of Computer Science VŠB-Technical University of Ostrava Ostrava Czech Republic Dept. of Comput. Sci. VSB-Tech. Univ. of Ostrava Ostrava Czech Republic

ISBN: (纸本)9781509021802

The search of k-nearest neighbors (K-NN) is a very common task. The K-NN is utilized in many algorithms and scientific areas like clustering, classification, machine learning, N-body simulation, triangulation, image processing and/or video processing. However, the naive implementation of the K-NN is very slow. There are many novel algorithms for the K-NN search, but they are usually based on the hierarchical clustering. The parallelization of those algorithms is a little bit tricky task. This paper primarily presents a novel parallel method for searching the k-nearest neighbors. An appropriate clustering of a sparse space based on the regular grid and the parallel search of the K-NN using the precomputed clusters are introduced. The whole method is designed for the parallel GPU computation and it is implemented on the CUDA architecture. The presented K-NN is utilized to speed up a force-directed graph layout algorithm, which can visually demonstrate the suitability of found neighbors, because they affect the layout quality. The graphs are widely used in social network analysis, computer networks or large information systems like photographic databases or multimedia databases to visualize relationships between elements. The achieved results and performance tests are presented as well.

关键词： Layout Graphics processing units Clustering algorithms Algorithm design and analysis Heuristic algorithms Acceleration Artificial neural networks

来源：评论

学校读者我要写书评

暂无评论

A Novel Illumination Compensation Method with Enhanced Retinex

A Novel Illumination Compensation Method with Enhanced Retin...

引用

International Conference on Information Science and Control Engineering (ICISCE)

作者： Libo Zhang Lin Yang Tiejian Luo Yihan Sun University of Chinese Academy of Sciences Beijing China

ISBN: (纸本)9781509025367

Illumination compensation is a typical way for increasing the performance of image processing. Retinex algorithm compensates illumination work properly, however, it cannot deal with color constancy of images appropriately. This paper proposes a novel illumination compensation method by converting the image from RGB color space to HSI color space, and with I component enhanced by Retinex algorithm and S component adjusted adaptively in accordance with correlation coefficients. Experimental results show that our method can not only maintain the color of images when the color space conversion from HIS to RGB, but also get better image quality with enhanced brightness, contrast and information. It has the trait of upholding the color constancy.

关键词： image color analysis Lighting Algorithm design and analysis Mathematical model Heuristic algorithms Brightness Visual systems

来源：评论

学校读者我要写书评

暂无评论

UFO — a scalable platform for high-speed synchrotron X-ray imaging

UFO — a scalable platform for high-speed synchrotron X-ray ...

引用

IEEE Symposium on Nuclear Science (NSS/MIC)

作者： Andreas Kopmann Suren Chilingaryan Matthias Vogelgesang Timo Dritschler Andrey Shkarin Roman Shkarin Tomy dos Santos Rolo Tomas Farago Thomas van de Kamp Matthias Balzer Michele Caselle Marc Weber Tilo Baumbach Karlsruhe Institute of Technology Germany

New imaging stations aim for high spatial and temporal resolution and are characterized by ever increasing sampling rates and demanding data processing workflows. Key to successful imaging experiments is to open up high-performance computing resources. This includes carefully selected components for computing hardware and development of advanced imaging algorithms optimized for efficient use of parallel processor architectures. We present the novel UFO computing platform for online data processing for imaging experiments and image-based feedback. The platform handles the full data life cycle from the X-ray detector to long-term data archives. Core components of this system are an FPGA platform for ultra-fast data acquisition, the GPU-based UFO image processing framework, and the fast control system “Concert”. Reconstruction algorithms implemented in the UFO framework are optimized for the latest GPU architectures and provide a reconstruction throughput in the GB/s-range. The control system “Concert” integrates high-speed computing nodes and fast beamline devices and thus enables image-based control loops and advanced workflow automation for efficient beam time usage. Low latencies are ensured by direct communication between FPGA and GPUs using AMDs DirectGMA technology. Time resolved tomography is supported by cutting edge regularization methods for high quality reconstructions with a reduced number of projections. The new infrastructure at ANKA has dramatically accelerated tomography from hours to second and resulted in new application fields, like high-throughput tomography, pump-probe radiography and stroboscopic tomography. Ultra-fast X-ray cine-tomography for the first time allows one to observe internal dynamics of moving millimeter-sized objects in real-time.

关键词： Cameras Tomography image reconstruction Control systems Graphics processing units Computer architecture

来源：评论

学校读者我要写书评

暂无评论

Low-rank plus sparse decomposition for exoplanet detection in direct-imaging ADI sequences The LLSG algorithm

引用

ASTRONOMY & ASTROPHYSICS 2016年第1期589卷

作者： Gonzalez, C. A. Gomez Absil, O. Absil, P. -A. Van Droogenbroeck, M. Mawet, D. Surdej, J. Univ Liege Inst Astrophys & Geophys Allee Six Aout 19c B-4000 Liege Belgium Catholic Univ Louvain Dept Math Engn B-1348 Louvain La Neuve Belgium Univ Liege Inst Montefiore B-4000 Liege Belgium CALTECH Dept Astron Pasadena CA 91125 USA

Context. Data processing constitutes a critical component of high-contrast exoplanet imaging. Its role is almost as important as the choice of a coronagraph or a wavefront control system, and it is intertwined with the chosen observing strategy. Among the data processing techniques for angular differential imaging (ADI), the most recent is the family of principal component analysis (PCA) based algorithms. It is a widely used statistical tool developed during the first half of the past century. PCA serves, in this case, as a subspace projection technique for constructing a reference point spread function (PSF) that can be subtracted from the science data for boosting the detectability of potential companions present in the data. Unfortunately, when building this reference PSF from the science data itself, PCA comes with certain limitations such as the sensitivity of the lower dimensional orthogonal subspace to non-Gaussian noise. Aims. Inspired by recent advances in machine learning algorithms such as robust PCA, we aim to propose a localized subspace projection technique that surpasses current PCA-based post-processing algorithms in terms of the detectability of companions at near real-time speed, a quality that will be useful for future direct imaging surveys. Methods. We used randomized low-rank approximation methods recently proposed in the machine learning literature, coupled with entry-wise thresholding to decompose an ADI image sequence locally into low-rank, sparse, and Gaussian noise components (LLSG). This local three-term decomposition separates the starlight and the associated speckle noise from the planetary signal, which mostly remains in the sparse term. We tested the performance of our new algorithm on a long ADI sequence obtained on beta Pictoris with VLT/NACO. Results. Compared to a standard PCA approach, LLSG decomposition reaches a higher signal-to-noise ratio and has an overall better performance in the receiver operating characteristic space. Th

关键词： methods: data analysis techniques: high angular resolution techniques: image processing planetary systems planets and satellites: detection

来源：评论

学校读者我要写书评

暂无评论

Fully automatic image colorization based on Convolutional Neural Network

Fully automatic image colorization based on Convolutional Ne...

引用

International Conference on Pattern Recognition

作者： Domonkos Varga Tamás Szirányi MTA SZTAKI Institute for Computer Science and Control Department of Networked Systems and Services Budapest University of Technology Department of Material Handling and Logistics Systems Budapest University of Technology

ISBN: (纸本)9781509048489

This paper deals with automatic image colorization. This is a very difficult task, since it is an ill-posed problem that usually requires user intervention to achieve high quality. A fully automatic approach is proposed that is able to produce realistic colorization of an input grayscale image. Motivated by the recent success of deep learning techniques in image processing, we propose a feed-forward, two-stage architecture based on Convolutional Neural Network that predicts the U and V color channels. Unlike most of the previous works, this paper presents a fully automatic colorization which is able to produce high-quality and realistic colorization even of complex scenes. Comprehensive experiments and qualitative and quantitative evaluations were conducted on the images of SUN database and on other images. We have found that Quaternion Structural Similarity (QSSIM) gives in some degree a good base for quantitative evaluation, that is why we chose QSSIM as an index-number for the quality of colorization.

关键词： image color analysis Semantics Feature extraction Gray-scale Neural networks Databases Prediction algorithms

来源：评论

学校读者我要写书评

暂无评论

Variational mode feature-based hyperspectral image classification

引用

Advances in Intelligent systems and Computing 2016年 380卷 365-373页

作者： Nechikkat, Nikitha Sowmya, V. Soman, K.P. Center for Excellence in Computational Engineering and Networking Amrita Vishwa Vidyapeetham CoimbatoreTamil Nadu India

ISBN: (纸本)9788132225225

Hyperspectral image analysis is considered as a promising technology in the field of remote sensing over the past decade. There are various processing and analysis techniques developed that interpret and extract the maximum information from high-dimensional hyperspectral datasets. The processing techniques significantly improve the performance of standard algorithms. This paper uses variational mode decomposition (VMD) as the processing algorithm for hyperspectral data scenarios followed by classification based on sparse representation. Variational Mode Decomposition decomposes the experimental data set into few different modes of separate spectral bands, which are unknown. These modes are given as raw input to the classifier for performance analysis. Orthogonal matching pursuit (OMP), the sparsity-based algorithm is used for classification. The proposed work is experimented on the standard dataset, namely Indian pines collected by the airborne visible/infrared imaging spectrometer (AVIRIS). The classification accuracy obtained on the hyperspectral data before and after applying Variational Mode Decomposition was analyzed. The experimental result shows that the proposed work leads to an improvement in the overall accuracy from 84.82 to 89.78%, average accuracy from 85.03 to 89.53% while using 40% data pixels for training. © Springer India 2016.

关键词： Variational mode decomposition

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：