Dictionary Representation Research Articles

Joint Estimation of Dictionary and Image from Compressive Samples

Effective compressed sensing of images relies on the prior knowledge of a well-suited dictionary for sparse representation of the target image. In the absence of such knowledge, which is a typical scenario in real-world applications, the following question arises: Would it be advantageous to take an off-the-shelf overcomplete dictionary and fine-tune the dictionary with respect to (the observed part of) the image? A primary obstacle in this approach is overfitting, i.e., the loss in model generalization to the whole image. In this paper, we establish that local dictionary optimization using the compressive samples reduces the image recovery error—relative to the off-the-shelf recovery—with an overwhelming probability that depends on the sampling matrix. We present joint estimation of dictionary and image (JEDI), an iterative algorithm for dictionary fine-tuning from compressive samples and analyze its performance for image recovery. Our algorithmic analysis is supplemented with numerical simulations under different random sampling patterns and off-the-shelf dictionary initializations.

Dictionary-based fiber orientation estimation with improved spatial consistency.

Diffusion magnetic resonance imaging (dMRI) has enabled in vivo investigation of white matter tracts. Fiber orientation (FO) estimation is a key step in tract reconstruction and has been a popular research topic in dMRI analysis. In particular, the sparsity assumption has been used in conjunction with a dictionary-based framework to achieve reliable FO estimation with a reduced number of gradient directions. Because image noise can have a deleterious effect on the accuracy of FO estimation, previous works have incorporated spatial consistency of FOs in the dictionary-based framework to improve the estimation. However, because FOs are only indirectly determined from the mixture fractions of dictionary atoms and not modeled as variables in the objective function, these methods do not incorporate FO smoothness directly, and their ability to produce smooth FOs could be limited. In this work, we propose an improvement to Fiber Orientation Reconstruction using Neighborhood Information (FORNI), which we call FORNI+; this method estimates FOs in a dictionary-based framework where FO smoothness is better enforced than in FORNI alone. We describe an objective function that explicitly models the actual FOs and the mixture fractions of dictionary atoms. Specifically, it consists of data fidelity between the observed signals and the signals represented by the dictionary, pairwise FO dissimilarity that encourages FO smoothness, and weighted ℓ1-norm terms that ensure the consistency between the actual FOs and the FO configuration suggested by the dictionary representation. The FOs and mixture fractions are then jointly estimated by minimizing the objective function using an iterative alternating optimization strategy. FORNI+ was evaluated on a simulation phantom, a physical phantom, and real brain dMRI data. In particular, in the real brain dMRI experiment, we have qualitatively and quantitatively evaluated the reproducibility of the proposed method. Results demonstrate that FORNI+ produces FOs with better quality compared with competing methods.

Localization of eye Saccadic signatures in Electrooculograms using sparse representations with data driven dictionaries

Abstract In this paper, we propose two methods for localizing saccadic eye movement signatures from Electrooculograms (EOG). The first approach uses a sparse representation of data-driven dictionaries of saccadic movements. In this approach, we match the EOG subsequence with the dictionary element using distance metrics to identify the saccades. The second approach is to compare a saccadic signature template with the EOG subsequence using Dynamic Time Warping (DTW). We find that the proposed methods have advantages over one another in context specific applications. The first method is significantly faster with considerable accuracy, while the second approach is more accurate.

Laplacian sparse dictionary learning for image classification based on sparse representation

Sparse representation is a mathematical model for data representation that has proved to be a powerful tool for solving problems in various fields such as pattern recognition, machine learning, and computer vision. As one of the building blocks of the sparse representation method, dictionary learning plays an important role in the minimization of the reconstruction error between the original signal and its sparse representation in the space of the learned dictionary. Although using training samples directly as dictionary bases can achieve good performance, the main drawback of this method is that it may result in a very large and inefficient dictionary due to noisy training instances. To obtain a smaller and more representative dictionary, in this paper, we propose an approach called Laplacian sparse dictionary (LSD) learning. Our method is based on manifold learning and double sparsity. We incorporate the Laplacian weighted graph in the sparse representation model and impose the l1-norm sparsity on the dictionary. An LSD is a sparse overcomplete dictionary that can preserve the intrinsic structure of the data and learn a smaller dictionary for each class. The learned LSD can be easily integrated into a classification framework based on sparse representation. We compare the proposed method with other methods using three benchmark-controlled face image databases, Extended Yale B, ORL, and AR, and one uncontrolled person image dataset, i-LIDS-MA. Results show the advantages of the proposed LSD algorithm over state-of-the-art sparse representation based classification methods.

Atmospheric inverse modeling via sparse reconstruction

Abstract. Many applications in atmospheric science involve ill-posed inverse problems. A crucial component of many inverse problems is the proper formulation of a priori knowledge about the unknown parameters. In most cases, this knowledge is expressed as a Gaussian prior. This formulation often performs well at capturing smoothed, large-scale processes but is often ill equipped to capture localized structures like large point sources or localized hot spots. Over the last decade, scientists from a diverse array of applied mathematics and engineering fields have developed sparse reconstruction techniques to identify localized structures. In this study, we present a new regularization approach for ill-posed inverse problems in atmospheric science. It is based on Tikhonov regularization with sparsity constraint and allows bounds on the parameters. We enforce sparsity using a dictionary representation system. We analyze its performance in an atmospheric inverse modeling scenario by estimating anthropogenic US methane (CH4) emissions from simulated atmospheric measurements. Different measures indicate that our sparse reconstruction approach is better able to capture large point sources or localized hot spots than other methods commonly used in atmospheric inversions. It captures the overall signal equally well but adds details on the grid scale. This feature can be of value for any inverse problem with point or spatially discrete sources. We show an example for source estimation of synthetic methane emissions from the Barnett shale formation.

Graph Regularized and Locality-Constrained Coding for Robust Visual Tracking

Visual tracking is complicated due to factors, such as occlusion, background clutter, abrupt target motion, and illumination variations, among others. In recent years, subspace representation and sparse coding techniques have demonstrated significant improvements in tracking. However, performance gain in tracking has been at the expense of losing locality and similarity attributes among the instances to be encoded. In this paper, a graph regularized and locality-constrained coding (GRLC) technique that encapsulates local manifold structure of the data in order to preserve locality and similarity information among instances is proposed. The GRLC methodology incorporates a similarity-preserving term within the objective function of the locality-constrained linear coding model, thereby overcoming some of the inherent instability issues common to such coding methods. In the proposed GRLC scheme, a graph Laplacian regularizer is chosen as a smoothing operator to learn both the representation dictionary and the coefficients by preserving the local structure of the data. This graph Laplacian smoothing operator ensures that the representations vary smoothly along the geodesics of the data manifold. Thus, by deriving the objective function of the GRLC method, a discriminative dictionary of instances can be iteratively obtained and the corresponding coefficients for each candidate can be computed using this learned dictionary. Finally, an effective observation likelihood function based on reconstruction error and a simple dictionary update scheme for visual target tracking are also proposed. Experimental results on the CVPR2013 visual tracker benchmark have demonstrated a favorable performance of the proposed technique both in terms of accuracy and robustness.

Training samples-optimizing based dictionary learning algorithm for MR sparse superresolution reconstruction

Abstract Magnetic Resonance (MR) imaging is widely used in diseases diagnosis. The hardware imaging arrives the limitation of resolution, and the high radiation intensity and time of magnetic hurts the human body. The software-based image super-resolution technology is prospective to solve the problem, especially with good excellent performance by sparse reconstruction-based image super-resolution. Dictionary generating is crucial issue of effecting the performance of the super-resolution algorithm, because of without considering the potential discriminative information during dictionary generating. For this problem, we propose the training samples-optimized dictionary learning algorithm for MR sparse super-resolution reconstruction. The gray-consistency & gradient joined diversity-based dictionary representation method is proposed to select the optimal images for the dictionary training. The dictionary training method is evaluated with the framework of sparse reconstruction-based MR imaging. Results show that the proposed dictionary selection framework is feasible and effective to improve the quality of sparse reconstruction-based MR super-resolution.

FUSION OF MULTI-SCALE DEMS FROM DESCENT AND NAVCM IMAGES OF CHANG’E-3 USING COMPRESSED SENSING METHOD

Abstract. The multi-source DEMs generated using the images acquired in the descent and landing phase and after landing contain supplementary information, and this makes it possible and beneficial to produce a higher-quality DEM through fusing the multi-scale DEMs. The proposed fusion method consists of three steps. First, source DEMs are split into small DEM patches, then the DEM patches are classified into a few groups by local density peaks clustering. Next, the grouped DEM patches are used for sub-dictionary learning by stochastic coordinate coding. The trained sub-dictionaries are combined into a dictionary for sparse representation. Finally, the simultaneous orthogonal matching pursuit (SOMP) algorithm is used to achieve sparse representation. We use the real DEMs generated from Chang’e-3 descent images and navigation camera (Navcam) stereo images to validate the proposed method. Through the experiments, we have reconstructed a seamless DEM with the highest resolution and the largest spatial coverage among the input data. The experimental results demonstrated the feasibility of the proposed method.

Estimation of white matter fiber parameters from compressed multiresolution diffusion MRI using sparse Bayesian learning

We present a sparse Bayesian unmixing algorithm BusineX: Bayesian Unmixing for Sparse Inference-based Estimation of Fiber Crossings (X), for estimation of white matter fiber parameters from compressed (under-sampled) diffusion MRI (dMRI) data. BusineX combines compressive sensing with linear unmixing and introduces sparsity to the previously proposed multiresolution data fusion algorithm RubiX, resulting in a method for improved reconstruction, especially from data with lower number of diffusion gradients. We formulate the estimation of fiber parameters as a sparse signal recovery problem and propose a linear unmixing framework with sparse Bayesian learning for the recovery of sparse signals, the fiber orientations and volume fractions. The data is modeled using a parametric spherical deconvolution approach and represented using a dictionary created with the exponential decay components along different possible diffusion directions. Volume fractions of fibers along these directions define the dictionary weights. The proposed sparse inference, which is based on the dictionary representation, considers the sparsity of fiber populations and exploits the spatial redundancy in data representation, thereby facilitating inference from under-sampled q-space. The algorithm improves parameter estimation from dMRI through data-dependent local learning of hyperparameters, at each voxel and for each possible fiber orientation, that moderate the strength of priors governing the parameter variances. Experimental results on synthetic and in-vivo data show improved accuracy with a lower uncertainty in fiber parameter estimates. BusineX resolves a higher number of second and third fiber crossings. For under-sampled data, the algorithm is also shown to produce more reliable estimates.

Fisher discrimination dictionary pair learning for image classification

Dictionary learning has played an important role in the success of sparse representation. Although several dictionary learning approaches have been developed for image classification, discriminative dictionary pair learning, i.e., jointly learning a synthesis dictionary and an analysis dictionary, is still in its infant stage. In this paper, we proposed a novel model of Fisher discrimination dictionary pair learning (FDDPL), in which Fisher discrimination information is embedded into analysis representation, analysis dictionary, and synthesis dictionary representation. With the proposed Fisher-like discrimination term, discrimination of both synthesis dictionary representation and analysis dictionary representation is introduced into the dictionary pair learning model. An iterative algorithm to efficiently solve the proposed FDDPL and a FDDPL based classifier are also presented in this paper. The experiments on face recognition, scene categorization, gender classification, and action recognition clearly show the advantage of the proposed FDDPL.

Dictionaries and distributions: Combining expert knowledge and large scale textual data content analysis : Distributed dictionary representation.

Theory-driven text analysis has made extensive use of psychological concept dictionaries, leading to a wide range of important results. These dictionaries have generally been applied through word count methods which have proven to be both simple and effective. In this paper, we introduce Distributed Dictionary Representations (DDR), a method that applies psychological dictionaries using semantic similarity rather than word counts. This allows for the measurement of the similarity between dictionaries and spans of text ranging from complete documents to individual words. We show how DDR enables dictionary authors to place greater emphasis on construct validity without sacrificing linguistic coverage. We further demonstrate the benefits of DDR on two real-world tasks and finally conduct an extensive study of the interaction between dictionary size and task performance. These studies allow us to examine how DDR and word count methods complement one another as tools for applying concept dictionaries and where each is best applied. Finally, we provide references to tools and resources to make this method both available and accessible to a broad psychological audience.

Слово «легенда» в речевом употреблении и в словарном отражении: заимствование, функционирование, идеология

Слово «легенда» в речевом употреблении и в словарном отражении: заимствование, функционирование, идеология

Noise-aware dictionary-learning-based sparse representation framework for detection and removal of single and combined noises from ECG signal.

Automatic electrocardiogram (ECG) signal enhancement has become a crucial pre-processing step in most ECG signal analysis applications. In this Letter, the authors propose an automated noise-aware dictionary learning-based generalised ECG signal enhancement framework which can automatically learn the dictionaries based on the ECG noise type for effective representation of ECG signal and noises, and can reduce the computational load of sparse representation-based ECG enhancement system. The proposed framework consists of noise detection and identification, noise-aware dictionary learning, sparse signal decomposition and reconstruction. The noise detection and identification is performed based on the moving average filter, first-order difference, and temporal features such as number of turning points, maximum absolute amplitude, zerocrossings, and autocorrelation features. The representation dictionary is learned based on the type of noise identified in the previous stage. The proposed framework is evaluated using noise-free and noisy ECG signals. Results demonstrate that the proposed method can significantly reduce computational load as compared with conventional dictionary learning-based ECG denoising approaches. Further, comparative results show that the method outperforms existing methods in automatically removing noises such as baseline wanders, power-line interference, muscle artefacts and their combinations without distorting the morphological content of local waves of ECG signal.

Progressive Compression of 3D Mesh Geometry Using Sparse Approximations from Redundant Frame Dictionaries

The proliferation of sophisticated techniques for 3D model acquisition and graphic design in recent years is resulting in a rapid increase in the number and complexity of 3D models being produced. At the same time, the need for exchanging 3D models is also rising due to new computing paradigms (e.g., cloud computing) and an increase in collaborative applications and online social networks. This has created a need for efficient compression techniques for 3D models, which are progressive, allow high quality or even lossless reconstruction, and minimize the amount of transferred data. Most of the research on 3D model compression to date has focused on the compression of triangular surface meshes, as this is the most common representation for 3D objects. In particular, progressive compression techniques for 3D mesh models have become very popular in recent years, as they allow a mesh to be reconstructed from coarse to fine levels of detail at the decoder while the bitstream of data is still being received. Currently the best-performing progressive compression techniques seem to be those that use transform coding ideas borrowed from traditional signal and image processing. The fundamental assumption behind such techniques is that the input data can be represented as some sparse linear combination of atoms taken from a representative dictionary of atoms that constitute the new representation domain. Traditionally, these dictionaries have been chosen to be orthogonal bases; however, recent research into alternative signal representations has shown that overcomplete dictionaries, or frames, may be able to produce even sparser representations than orthogonal bases can. In the field of 3D mesh compression, the idea of using redundant representations has only just started to surface. The main challenge here is currently the creation of a good overcomplete dictionary, particularly for mesh models with irregular connectivity that needs to be preserved. In light of this, the main objective of the work in this thesis is to develop a new method for constructing a frame, which can be applied to a manifold triangle mesh directly (i.e., with no requirement for a prior remeshing to a more regular domain) and can be used to obtain sparse approximations of the mesh geometry in a progressive compression scenario. This is the main contribution of this thesis. The frames that we propose are constructed from the unit-norm eigenvectors of a combinatorial mesh Laplacian matrix, plus a large number of redundant linear combinations of these eigenvectors. A sparse synthesis of the input mesh geometry is achieved by selecting atoms from the frame using the Matching Pursuit algorithm. The proposed frames can be applied directly to a manifold mesh with arbitrary topology and connectivity type, as long as the mesh can be represented as a connected graph. Our experimental results for a variety of different mesh models indicate that, for a given reconstruction quality, a sparser approximation of the input mesh geometry can be obtained with the proposed…

基于联合多特征字典稀疏表示的步态识别算法

现有的步态识别算法多采用模型特征或整体特征进行单一特征提取，在多视角等实际情况中算法鲁棒性较差、识别率较低。针对这一问题，本文提出了一种基于联合多特征字典稀疏表示的步态识别算法。该算法选择三种不同粒度的特征：均值形状PMS、步态能量图GEI与自建特征-区域面积序列RAS，构建特征训练字典并对特征样本进行多任务联合稀疏表示，最后通过计算最小累计残差得到测试样本类别，实现特征层融合。实验结果表明，相比单一特征提取与识别，所采用的多特征联合识别方法识别率更高，且在多视角下具有一定鲁棒性，实现了特征之间的信息互补。 Most of the existing gait recognition algorithms extract the single feature using model features or global features. However, these algorithms usually have a poor robustness and a low recognition rate in practical situations such as multi-angle. To solve this problem, a gait recognition algorithm based on joint sparse representation of multi-feature dictionaries is proposed in this paper. In this algorithm, three characteristics in different particle size are selected: Procrustes Mean Shape, Gait Energy Image and Region Area Sequence which is structured in this article. Feature training dictionaries are constructed and a multidisciplinary sparse representation to feature samples is done. Finally, the test sample category is obtained by calculating the minimum cumulative residual and achieves the integration of feature layer. Experimental results show that the multi-feature joint recognition method used in this paper has a higher recognition rate and a certain robustness at multiple angles compared to single feature extraction and recognition. This paper basically fulfills the complementary information between features.

A platform-independent method to reduce CT truncation artifacts using discriminative dictionary representations.

When the scan field of view (SFOV) of a CT system is not large enough to enclose the entire cross-section of the patient, or the patient needs to be positioned partially outside the SFOV for certain clinical applications, truncation artifacts often appear in the reconstructed CT images. Many truncation artifact correction methods perform extrapolations of the truncated projection data based on certain a priori assumptions. The purpose of this work was to develop a novel CT truncation artifact reduction method that directly operates on DICOM images. The blooming of pixel values associated with truncation was modeled using exponential decay functions, and based on this model, a discriminative dictionary was constructed to represent truncation artifacts and nonartifact image information in a mutually exclusive way. The discriminative dictionary consists of a truncation artifact subdictionary and a nonartifact subdictionary. The truncation artifact subdictionary contains 1000 atoms with different decay parameters, while the nonartifact subdictionary contains 1000 independent realizations of Gaussian white noise that are exclusive with the artifact features. By sparsely representing an artifact-contaminated CT image with this discriminative dictionary, the image was separated into a truncation artifact-dominated image and a complementary image with reduced truncation artifacts. The artifact-dominated image was then subtracted from the original image with an appropriate weighting coefficient to generate the final image with reduced artifacts. This proposed method was validated via physical phantom studies and retrospective human subject studies. Quantitative image evaluation metrics including the relative root-mean-square error (rRMSE) and the universal image quality index (UQI) were used to quantify the performance of the algorithm. For both phantom and human subject studies, truncation artifacts at the peripheral region of the SFOV were effectively reduced, revealing soft tissue and bony structure once buried in the truncation artifacts. For the phantom study, the proposed method reduced the relative RMSE from 15% (original images) to 11%, and improved the UQI from 0.34 to 0.80. A discriminative dictionary representation method was developed to mitigate CT truncation artifacts directly in the DICOM image domain. Both phantom and human subject studies demonstrated that the proposed method can effectively reduce truncation artifacts without access to projection data.

A Perturbation-Based Approach for Compressed Sensing Radar Imaging

This letter is concerned with two-dimensional radar imaging via compressed sensing. First, a perturbation-based sparse representation dictionary is established to alleviate the basis mismatch effect, reduce the memory requirement, and accelerate the imaging process. Then, an iterative algorithm based on variational Bayesian inference is developed for sparse recovery that is user-parameter-free and suitable for imposing a priori constraint. Experimental results based on both synthetic and real data verify the performance of the proposed approach.

Joint and collaborative representation with local adaptive convolution feature for face recognition with single sample per person

With the aid of a universal facial variation dictionary, sparse representation based classifier (SRC) has been naturally extended for face recognition (FR) with single sample per person (SSPP) and achieved promising performance. However, extracting discriminative facial features and building powerful representation framework for classifying query face images are still the bottlenecks of improving the performance of FR with SSPP. In this paper, by densely sampling and sparsely detecting facial points, we extract complete and robust local regions and learn convolution features adaptive to the local regions and discriminative to the face identity by using convolutional neural networks (CNN). With this powerful facial description and a generic face dataset with common facial variations, a joint and collaborative representation framework, which performs representation for each local region of the query face image while requires all regions of the query face image to have similar representation coefficients, is presented to exploit the distinctiveness and commonality of different local regions. In the proposed joint and collaborative representation with local adaptive convolution feature (JCR-ACF), both discriminative local facial features that are robust to various facial variations and powerful representation dictionaries of facial variations that can overcome the small-sample-size problem are fully exploited. JCR-ACF has been extensively evaluated on several popular databases including AR, CMU Multi-PIE, LFW and the large-scale CASIA-WebFace databases. Experimental results demonstrate the much higher robustness and effectiveness of JCR-ACF to complex facial variations compared to the state-of-the-art methods.

A Novel Multi-Focus Image Fusion Method Based on Stochastic Coordinate Coding and Local Density Peaks Clustering

The multi-focus image fusion method is used in image processing to generate all-focus images that have large depth of field (DOF) based on original multi-focus images. Different approaches have been used in the spatial and transform domain to fuse multi-focus images. As one of the most popular image processing methods, dictionary-learning-based spare representation achieves great performance in multi-focus image fusion. Most of the existing dictionary-learning-based multi-focus image fusion methods directly use the whole source images for dictionary learning. However, it incurs a high error rate and high computation cost in dictionary learning process by using the whole source images. This paper proposes a novel stochastic coordinate coding-based image fusion framework integrated with local density peaks. The proposed multi-focus image fusion method consists of three steps. First, source images are split into small image patches, then the split image patches are classified into a few groups by local density peaks clustering. Next, the grouped image patches are used for sub-dictionary learning by stochastic coordinate coding. The trained sub-dictionaries are combined into a dictionary for sparse representation. Finally, the simultaneous orthogonal matching pursuit (SOMP) algorithm is used to carry out sparse representation. After the three steps, the obtained sparse coefficients are fused following the max L1-norm rule. The fused coefficients are inversely transformed to an image by using the learned dictionary. The results and analyses of comparison experiments demonstrate that fused images of the proposed method have higher qualities than existing state-of-the-art methods.

Multimodal Sentiment Intensity Analysis in Videos: Facial Gestures and Verbal Messages

People share their opinions, stories, and reviews through online video sharing websites every day. The automatic analysis of these online opinion videos is bringing new or understudied research challenges to the field of computational linguistics and multimodal analysis. Among these challenges is the fundamental question of exploiting the dynamics between visual gestures and verbal messages to be able to better model sentiment. This article addresses this question in four ways: introducing the first multimodal dataset with opinion-level sentiment intensity annotations; studying the prototypical interaction patterns between facial gestures and spoken words when inferring sentiment intensity; proposing a new computational representation, called multimodal dictionary, based on a language-gesture study; and evaluating the authors' proposed approach in a speaker-independent paradigm for sentiment intensity prediction. The authors' study identifies four interaction types between facial gestures and verbal content: neutral, emphasizer, positive, and negative interactions. Experiments show statistically significant improvement when using multimodal dictionary representation over the conventional early fusion representation (that is, feature concatenation).