检索结果-内蒙古大学图书馆

26th IEEE Annual Canadian Conference on Electrical and Computer Engineering (CCECE)

作者： Tharmalingam, Mathiruban Raahemifar, Kaamran Ryerson Univ Dept Elect Engn & Comp Eng Toronto ON Canada

ISBN: (纸本)9781479900336;9781479900312

Complex signals such as images, audio and video recordings can be represented by a large over-complete dictionary without significant compromise on the representation quality. An over-complete dictionary has many more columns than the number of rows. Large over-complete dictionaries can produce sparse representation vectors and provide significant improvements in the reconstructed signal quality because it contains many patterns to select from. The use of the over-complete dictionaries and sparse coding has been successfully applied in compression, de-noising, and pattern recognition applications within the last few decades. An example of an over-complete dictionary that has seen a great deal of success in image processing applications is the Discrete Cosine Transform (DCT) dictionary. However, we propose a novel non-linear over-complete dictionary that improves the quality of the signal representation while reducing the number of non-zero elements to represent the signal. The proposed non-linear dictionary has demonstrated through experimental results to be superior to the DCT dictionary by achieving higher signal to noise ratio (SNR) in the reconstructed images.

关键词： sparse coding Non-linear dictionaries atoms DCT time-shifted OMP image reconstruction

来源：评论

学校读者我要写书评

暂无评论

CLASSIFICATION OF TUMOR HISTOPATHOLOGY VIA sparse FEATURE LEARNING

CLASSIFICATION OF TUMOR HISTOPATHOLOGY VIA SPARSE FEATURE LE...

引用

IEEE 10th International Symposium on Biomedical Imaging - From Nano to Macro (ISBI)

作者： Nayak, Nandita Chang, Hang Borowsky, Alexander Spellman, Paul Parvin, Bahram Univ Calif Berkeley Lawrence Berkeley Natl Lab Div Life Sci Berkeley CA 94720 USA Univ Calif Berkeley Ctr Comparat Med Berkeley CA 94720 USA Oregon Hlth & Sci Univ Ctr Spatial Syst Biomed Portland OR USA

ISBN: (纸本)9781467364553

Our goal is to decompose whole slide images (WSI) of histology sections into distinct patches (e. g., viable tumor, necrosis) so that statistics of distinct histopathology can be linked with the outcome. Such an analysis requires a large cohort of histology sections that may originate from different laboratories, which may not use the same protocol in sample preparation. We have evaluated a method based on a variation of the restricted Boltzmann machine (RBM) that learns intrinsic features of the image signature in an unsupervised fashion. Computed code, from the learned representation, is then utilized to classify patches from a curated library of images. The system has been evaluated against a dataset of small image blocks of 1k-by-1k that have been extracted from glioblastoma multiforme (GBM) and clear cell kidney carcinoma (KIRC) from the cancer genome atlas (TCGA) archive. The learned model is then projected on each whole slide image (e. g., of size 20k-by-20k pixels or larger) for characterizing and visualizing tumor architecture. In the case of GBM, each WSI is decomposed into necrotic, transition into necrosis, and viable. In the case of the KIRC, each WSI is decomposed into tumor types, stroma, normal, and others. Evaluation of 1400 and 2500 samples of GBM and KIRC indicates a performance of 84% and 81%, respectively.

关键词： Tumor characterization whole slide imaging feature learning sparse coding

来源：评论

学校读者我要写书评

暂无评论

The Application of sparse coding Algorithm Based on Kurtosis Criterion in Natural Image Compression

The Application of Sparse Coding Algorithm Based on Kurtosis...

引用

Chinese Conference on Pattern Recognition/1st CJK Joint Workshop on Pattern Recognition

作者： Shang, Li Huai, Wen-jun Suzhou Vocat Univ Elect & Informat Engn Dept Suzhou 215104 Peoples R China

ISBN: (纸本)9781424441990

This paper mainly discussed the algorithm of sparse coding based on the kurtosis criterion and its application in natural image compression. sparse coding of natural images is in fact a transformation coding method, and it can efficiently perform extracting natural images' features and compressing images. Utilizing the kurtosis as the punitive function of sparse coding method's sparsity, it can ensure both the sparsity and independence of feature coefficients of natural images, and extract more efficiently the edge features of images. Compared with the image compression methods of standard independent component analysis (ICA) and discrete cosine transfer (DCT), the simulation results show that our method proposed excels in natural image compression.

关键词： sparse coding kurtosis feature extraction compression

来源：评论

学校读者我要写书评

暂无评论

A NONLINEAR DICTIONARY FOR IMAGE RECONSTRUCTION

A NONLINEAR DICTIONARY FOR IMAGE RECONSTRUCTION

引用

IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Tharmalingam, Mathiruban Raahemifar, Kaamran Ryerson Univ Dept Elect Engn & Comp Eng Toronto ON Canada

ISBN: (纸本)9781479903566

Complex signals such as images, audio and video recordings can be represented by a large over complete dictionary without distinguishable compromise on the representation quality. Large over complete dictionaries with more patterns can be used to increase the sparse coding as well as provide significant improvements in signal representation quality. The use of the over-complete dictionaries and sparse coding has been successfully applied in compression, de-noising, and pattern recognition applications within the last few decades. One particular dictionary, the Discrete Cosine Transform (DCT) dictionary has seen a great deal of success in image processing applications. However, we propose a novel non-linear over-complete dictionary that is sparser than the DCT dictionary while improving the quality of the signal representation. The proposed non-linear dictionary has demonstrated through experimental results to be superior to the DCT dictionary by achieving higher signal to noise ratio (SNR) in the reconstructed images.

关键词： sparse coding Non-linear dictionaries atoms DCT dictionary image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Online Robust Object Tracking via A Sample-based Dynamic Dictionary

Online Robust Object Tracking via A Sample-based Dynamic Dic...

引用

6th International Congress on Image and Signal Processing (CISP)

作者： Liu, Yang Li, Yibo Nanjing Univ Aeronaut & Astronaut Coll Automat Engn Nanjing Jiangsu Peoples R China Shenyang Aerosp Univ Coll Automat Shenyang Peoples R China

ISBN: (纸本)9781479927630

We propose an online robust object tracking algorithm based on a sample-based dictionary. The sample-based dictionary in our method means that the over-completely dictionary of sparse coding algorithm is formed by using the sample basis extracted from video images. Different from the other tracking methods that use the object features and a set of boosted classifiers, the proposed algorithm considers the raw image patches around the object as basis vectors and the maximum a posteriori is used to decide the object position in the next frame. Our method is simple for the dictionary that is updated automatically and the object is updated to alleviate the visual drift problem in every frame without learning process, which needs much time consuming. Experiments are conducted on the challenging sequences to demonstrate that the proposed method is fast and effective. The results show that the proposed method outperforms the current state-of-the-art methods.

关键词： Object tracking sparse coding Bayesian tracking framework sample-based dynamic dictionary

来源：评论

学校读者我要写书评

暂无评论

IMAGE REGISTRATION BASED ON DYNAMIC DIRECTED GRAPHS WITH GROUPWISE IMAGE SIMILARITY

IMAGE REGISTRATION BASED ON DYNAMIC DIRECTED GRAPHS WITH GRO...

引用

IEEE 10th International Symposium on Biomedical Imaging - From Nano to Macro (ISBI)

作者： Tang, Zhenyu Jiang, Di Fan, Yong Chinese Acad Sci Inst Automat Natl Lab Pattern Recognit Beijing 100190 Peoples R China

ISBN: (纸本)9781467364553

We propose a groupwise image registration method using sparse coding and graph theoretic techniques. A sparse coding method is used to estimate image similarity measures among images to registered, yielding asymmetric, groupwise image similarity measures for each image to other images in the group. Based on the asymmetric groupwise image similarity measures among different images, a directed graph is built for learning a manifold of images so that a group center image can be identified and each image can be registered to the center image following a directed shortest path which decomposes a large deformation between two images into a series of small and anatomically meaningful deformations. The groupwise image similarity measures are progressively updated, so does the directed graph, facilitating accurate alignment of a group of images following directed shortest paths derived from the dynamic directed graphs. Compared with state of the art graph based registration methods, our method is more robust and able to achieve better registration accuracy.

关键词： Groupwise registration groupwise similarity dynamic graph sparse coding

来源：评论

学校读者我要写书评

暂无评论

Similarity Search in Visual Data

Similarity Search in Visual Data

引用

作者： Cherian, Anoop University of Minnesota

学位级别：Ph.D.

Contemporary times have witnessed a significant increase in the amount of data available on the Internet. Organizing such big data so that it is easily and promptly accessible, is a necessity that has been growing in importance. Among the various data modalities such as text, audio, etc., visual data (in the form of images and videos) constitute a major share of this available content. Contrary to other data modalities, visual data pose several significant challenges to storage and retrieval, namely (i) choosing an appropriate representation that can capture the essence of visual data is often non-trivial, and (ii) visual search and retrieval are often subjective, as a result computing semantically meaningful results is hard. On the other hand, visual data possesses rich structure. Exploiting this structure might help address these challenges. Motivated by these observations, this thesis explores new algorithms for efficient similarity search in structured visual data; “structure” is synonymous with the mathematical representation that captures desirable data properties. We will deal with two classes of such structures that are common in computer vision, namely (i) symmetric positive definite matrices as covariances, and (ii) sparse data representations in a dictionary learned from the data. Covariance valued data has found immense success in several mainstream computer vision applications such as visual surveillance, emotion recognition, face recognition, etc. Moreover, it is of fundamental importance in several other disciplines such as magnetic resonance imaging, speech recognition, etc. A technical challenge in computing similarities on such matrix valued data is their non-Euclidean nature. These matrices belong to a curved manifold where distances between data points are no more along straight lines, but along curved geodesics. As a result, state-of-the-art measures for comparing covariances tend to be slow. To address this issue, we propose a novel similarity

关键词： Covariance matrices Dictionary learning Dirichlet process Jensen-bregman logdet divergence Nearest neighbors sparse coding

来源：评论

学校读者我要写书评

暂无评论

Efficient occlusive components analysis

The Journal of Machine Learning Research

引用

The Journal of Machine Learning Research 2014年第1期15卷

作者： Kevin Murphy Bernhard Schölkopf Marc Henniges Richard E. Turner Maneesh Sahani Julian Eggert Jög Lücke Google Frankfurt Institute for Advanced Studies Goethe-University Frankfurt Frankfurt Germany Department of Engineering University of Cambridge Cambridge UK University College London London UK Honda Research Institute Europe GmbH Germany Cluster of Excellence Hearing4all and Faculty VI University of Oldenburg Oldenburg Germany and Frankfurt Institute for Advanced Studies Goethe-University Frankfurt Frankfurt Germany and Electrical Engineering and Computer Science Technical University Berlin Berlin Germany

We study unsupervised learning in a probabilistic generative model for occlusion. The model uses two types of latent variables: one indicates which objects are present in the image, and the other how they are ordered in depth. This depth order then determines how the positions and appearances of the objects present, specified in the model parameters, combine to form the image. We show that the object parameters can be learned from an unlabeled set of images in which objects occlude one another. Exact maximum-likelihood learning is intractable. Tractable approximations can be derived, however, by applying a truncated variational approach to Expectation Maximization (EM). In numerical experiments it is shown that these approximations recover the underlying set of object parameters including data noise and sparsity. Experiments on a novel version of the bars test using colored bars, and experiments on more realistic data, show that the algorithm performs well in extracting the generating components. The studied approach demonstrates that the multiple-causes generative approach can be generalized to extract occluding components, which links research on occlusion to the field of sparse coding approaches.

关键词： expectation truncation generative models occlusion sparse coding unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

Robust active binocular vision through intrinsically motivated learning

引用

FRONTIERS IN NEUROROBOTICS 2013年第2013期7卷 20-20页

作者： Lonini, Luca Forestier, Sebastien Teuliere, Celine Zhao, Yu Shi, Bertram E. Triesch, Jochen Goethe Univ Frankfurt Frankfurt Inst Adv Studies D-60438 Frankfurt Germany Ecole Normale Super Cachan Bretagne Bruz France Hong Kong Univ Sci & Technol Dept Elect & Comp Engn Hong Kong Hong Kong Peoples R China

The efficient coding hypothesis posits that sensory systems of animals strive to encode sensory signals efficiently by taking into account the redundancies in them. This principle has been very successful in explaining response properties of visual sensory neurons as adaptations to the statistics of natural images. Recently, we have begun to extend the efficient coding hypothesis to active perception through a form of intrinsically motivated learning: a sensory model learns an efficient code for the sensory signals while a reinforcement learner generates movements of the sense organs to improve the encoding of the signals. To this end, it receives an intrinsically generated reinforcement signal indicating how well the sensory model encodes the data. This approach has been tested in the context of binocular vison, leading to the autonomous development of disparity tuning and vergence control. Here we systematically investigate the robustness of the new approach in the context of a binocular vision system implemented on a robot. Robustness is an important aspect that reflects the ability of the system to deal with unmodeled disturbances or events, such as insults to the system that displace the stereo cameras. To demonstrate the robustness of our method and its ability to self-calibrate, we introduce various perturbations and test if and how the system recovers from them. We find that (1) the system can fully recover from a perturbation that can be compensated through the system's motor degrees of freedom, (2) performance degrades gracefully if the system cannot use its motor degrees of freedom to compensate for the perturbation, and (3) recovery from a perturbation is improved if both the sensory encoding and the behavior policy can adapt to the perturbation. Overall, this work demonstrates that our intrinsically motivated learning approach for efficient coding in active perception gives rise to a self-calibrating perceptual system of high robustness.

关键词： active perception sparse coding reinforcement learning robotics stereo vision vergence robustness

来源：评论

学校读者我要写书评

暂无评论

sparse Hashing for Fast Multimedia Search

引用

ACM TRANSACTIONS ON INFORMATION SYSTEMS 2013年第2期31卷 9-9页

作者： Zhu, Xiaofeng Huang, Zi Cheng, Hong Cui, Jiangtao Shen, Heng Tao Univ Queensland Sch Informat Technol & Elect Engn Brisbane Qld 4072 Australia Chinese Univ Hong Kong Dept Syst Engn & Engn Management Hong Kong Hong Kong Peoples R China Xidian Univ Sch Comp Sci & Technol Xian Peoples R China

Hash-based methods achieve fast similarity search by representing high-dimensional data with compact binary codes. However, both generating binary codes and encoding unseen data effectively and efficiently remain very challenging tasks. In this article, we focus on these tasks to implement approximate similarity search by proposing a novel hash based method named sparse hashing (SH for short). To generate interpretable (or semantically meaningful) binary codes, the proposed SH first converts original data into low-dimensional data through a novel nonnegative sparse coding method. SH then converts the low-dimensional data into Hamming space (i.e., binary encoding low-dimensional data) by a new binarization rule. After this, training data are represented by generated binary codes. To efficiently and effectively encode unseen data, SH learns hash functions by taking a-priori knowledge into account, such as implicit group effect of the features in training data, and the correlations between original space and the learned Hamming space. SH is able to perform fast approximate similarity search by efficient bit XOR operations in the memory of a modern PC with short binary code representations. Experimental results show that the proposed SH significantly outperforms state-of-the-art techniques.

关键词： Design Algorithms Performance Hashing indexing sparse coding multimedia search

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：