Sparse Feature Representation Research Articles

Building Recognition Based on Sparse Representation of Spatial Texture and Color Features

In this paper, we presented a novel building recognition method based on a sparse representation of spatial texture and color features. At present, the most popular methods are based on gist features, which can only roughly reflect the spatial information of building images. The proposed method, in contrast, uses multi-scale neighborhood sensitive histograms of oriented gradient (MNSHOGs) and color auto-correlogram (CA) to extract texture and color features of building images. Both the MNSHOG and the CA take spatial information of building images into account while calculating texture and color features. Then, color and texture features are combined to form joint features whose sparse representation can be dimensionally reduced by an autoencoder. Finally, an extreme learning machine is used to classify the combined features after dimensionality reduction into different classes. Several experiments were conducted on the benchmark Sheffield building dataset. The mean recognition rate of our method is much higher than that of the existing methods, which shows the effectiveness of our method.

Simultaneous Human Health Monitoring and Time-Frequency Sparse Representation Using EEG and ECG Signals

In the field of human health monitoring, intelligent diagnostic methods have drawn much attention recently to tackle the health problems and challenges faced by patients. In this paper, an efficient and flexible diagnostic method is proposed, which enables the simultaneous use of a machine learning method and sparsity-based representation technique. Specifically, the proposed method is based on a convolutional neural network (CNN) and generalized minimax-concave (GMC) method. First, measured potential signals, for instance, electroencephalogram (EEG) and electrocardiogram (ECG) signals are directly inputted into the designed network based on CNN for health conditions classification. The designed network adopts small convolution kernels to enhance the performance of feature extraction. In the training process, small batch samples are applied to improve the generalization of the model. Meanwhile, the “Dropout” strategy is applied to overcome the overfitting problem in fully connected layers. Then, for a record of the interested EEG or ECG signal, the sparse representation of useful time-frequency features can be estimated via the GMC method. Case studies of seizure detection and arrhythmia signal analysis are adopted to verify the performance of the proposed method. The experimental results demonstrate that the proposed method can effectively identify different health conditions and maximally enhance the sparsity of time-frequency features.

3D Patch-Based Sparse Learning for Style Feature Extraction

How to extract features of different shapes for style similarity evaluation is a very challenging research topic. Different from methods based on predefined style templates, we perform sparse learning and triplet embedding directly from the 3D shape local descriptors. The proposed method can adaptively generate element-level style features of different 3D patches for style similarity evaluation. The proposed algorithm mainly consists of the following steps. First, given a heterogeneous 3D shape collection, we learn mid-level style features by Siamese networks based on curvature-guided sampling directly on 3D shapes, which is a view-independent and part-independent approach. Second, we perform sparse coding to reduce the dimension of mid-level local features and achieve more compact and discriminative style features, named element-level features. Finally, we enhance our element-level style features by triplet embedding learning over crowdsourced triplets, which maps similar examples close to each other and dissimilar examples farther apart. The proposed algorithm can aggregate different local features into a global style feature representation and shows satisfactory discriminability. Our method can avoid the loss of style features from the 3D model to the views and can learn more discriminative style features by sparse learning. The experimental results show that our sparse style feature representation significantly improves the accuracy compared with recent state-of-the-art algorithms.

A Deep Neural Network Model using Random Forest to Extract Feature Representation for Gene Expression Data Classification

In predictive model development, gene expression data is associated with the unique challenge that the number of samples (n) is much smaller than the amount of features (p). This “n ≪ p” property has prevented classification of gene expression data from deep learning techniques, which have been proved powerful under “n > p” scenarios in other application fields, such as image classification. Further, the sparsity of effective features with unknown correlation structures in gene expression profiles brings more challenges for classification tasks. To tackle these problems, we propose a newly developed classifier named Forest Deep Neural Network (fDNN), to integrate the deep neural network architecture with a supervised forest feature detector. Using this built-in feature detector, the method is able to learn sparse feature representations and feed the representations into a neural network to mitigate the overfitting problem. Simulation experiments and real data analyses using two RNA-seq expression datasets are conducted to evaluate fDNN’s capability. The method is demonstrated a useful addition to current predictive models with better classification performance and more meaningful selected features compared to ordinary random forests and deep neural networks.

Halo-Free Multi-Exposure Image Fusion Based on Sparse Representation of Gradient Features

Due to sharp changes in local brightness in high dynamic range scenes, fused images obtained by the traditional multi-exposure fusion methods usually have an unnatural appearance resulting from halo artifacts. In this paper, we propose a halo-free multi-exposure fusion method based on sparse representation of gradient features for high dynamic range imaging. First, we analyze the cause of halo artifacts. Since the range of local brightness changes in high dynamic scenes may be far wider than the dynamic range of an ordinary camera, there are some invalid, large-amplitude gradients in the multi-exposure source images, so halo artifacts are produced in the fused image. Subsequently, by analyzing the significance of the local sparse coefficient in a luminance gradient map, we construct a local gradient sparse descriptor to extract local details of source images. Then, as an activity level measurement in the fusion method, the local gradient sparse descriptor is used to extract image features and remove halo artifacts when the source images have sharp local changes in brightness. Experimental results show that the proposed method obtains state-of-the-art performance in subjective and objective evaluation, particularly in terms of effectively eliminating halo artifacts.

Self-Tuned Descriptive Document Clustering Using a Predictive Network

Descriptive clustering consists of automatically organizing data instances into clusters and generating a descriptive summary for each cluster. The description should inform a user about the contents of each cluster without further examination of the specific instances, enabling a user to rapidly scan for relevant clusters. Selection of descriptions often relies on heuristic criteria. We model descriptive clustering as an auto-encoder network that predicts features from cluster assignments and predicts cluster assignments from a subset of features. The subset of features used for predicting a cluster serves as its description. For text documents, the occurrence or count of words, phrases, or other attributes provides a sparse feature representation with interpretable feature labels. In the proposed network, cluster predictions are made using logistic regression models, and feature predictions rely on logistic or multinomial regression models. Optimizing these models leads to a completely self-tuned descriptive clustering approach that automatically selects the number of clusters and the number of features for each cluster. We applied the methodology to a variety of short text documents and showed that the selected clustering, as evidenced by the selected feature subsets, are associated with a meaningful topical organization.

Dictionary Learning Inspired Deep Network for Scene Recognition

Scene recognition remains one of the most challenging problems in image understanding. With the help of fully connected layers (FCL) and rectified linear units (ReLu), deep networks can extract the moderately sparse and discriminative feature representation required for scene recognition. However, few methods consider exploiting a sparsity model for learning the feature representation in order to provide enhanced discriminative capability. In this paper, we replace the conventional FCL and ReLu with a new dictionary learning layer, that is composed of a finite number of recurrent units to simultaneously enhance the sparse representation and discriminative abilities of features via the determination of optimal dictionaries. In addition, with the help of the structure of the dictionary, we propose a new label discriminative regressor to boost the discrimination ability. We also propose new constraints to prevent overfitting by incorporating the advantage of the Mahalanobis and Euclidean distances to balance the recognition accuracy and generalization performance. Our proposed approach is evaluated using various scene datasets and shows superior performance to many state-of-the-art approaches.

DEMIAL: an active learning framework for multiple instance image classification using dictionary ensembles

In many machine-learning applications, each data point can be represented as a set of instances that create multiple instance learning (MIL) problems. Due to the structure of images, different regions can be interpreted as instances. Thus, multiple instances can be obtained for each image, which makes image categorization a MIL problem. With abundant unlabeled image data, this MIL problem can be solved using active learning algorithms. Active learning is a framework that utilizes unlabeled data in which labeling samples is a labor-intensive and expensive task. Although many effective MIL active learning methods have been developed, most of the existing algorithms do not take into account classifier and feature representation. In this work, we develop DEMIAL (dictionary ensembles multiple instance active learning), a multiple instance active learning method that utilizes sparse feature representation and classifier ensemble techniques. In the proposed active learning framework, we employ dictionary learning and compare uncertainty- and entropy-based instance selection techniques. Experimental results show that classifier ensembles benefit from active learning and the DEMIAL algorithm outperforms the kernel-based multiple instance active learning framework.

A Homotopy Iterative Hard Thresholding Algorithm With Extreme Learning Machine for Scene Recognition

Vision-based scene recognition aims to find a semantic explanation of a scene image. First, how to learn a sparse feature representation by mapping each cluster of patches into few dimensions is one challenging issue. Sparse coding techniques with $l_{1}$ norm regularization are widely used to learn sparse features. However, $l_{1}$ norm regularization may not achieve enough sparsity. Furthermore, the initial value has a great influence on the sparsity of codes during the iterative optimization process. Second, how to learn a classifier in a supervised manner in order to improve the generalization is another issue. This paper therefore proposes a scene recognition method and it mainly includes two processes. One is the homotopy iterative hard thresholding (HIHT) algorithm that encodes the sparse representations of local patches by incorporating an $l_{0}$ norm regularization. Furthermore, a homotopy continuity strategy is used to improve the sparsity of feature codes by adaptively tuning the regularization factor from large to small values and using the sparse solution of last iteration as the warm start of the next iteration. The other is the extreme learning machine (ELM)-based classifier. Experimental results in 15-class scene data set have shown that the HIHT algorithm outperforms other unsupervised sparse feature learning algorithms in terms of sparsity and entropy. Meanwhile, the ELM-based scene recognition method outperforms the other state-of-the art methods in terms of recognition accuracy.

Big data analytics enabled by feature extraction based on partial independence

Abstract Complex cells in primary visual cortex (V1) selectively respond to bars and edges at a particular location and orientation. Namely, they are relatively invariant to the phase as well as selective to the frequency and orientation emerging from natural images that are analogous to the characteristics of complex cells in V1 with the energy function of receptive fields (RFs) from tuning curve test with sinusoidal function in our related jobs. In this paper, we propose a feature learning algorithm based on the overcomplete AISA to apply on big data in parallel computing. In order to demonstrate the effectiveness of the overcomplete AISA features in the classification task, two feature representation architectures are evolved into the partial independent signal bases and partial independent factorial representation, respectively. Experiments on four datasets (Coil20, Extended YaleB, USPS, PIE), acquired conjunction with two classification architectures based on the overcomplete AISA features, show that the classification accuracy is mostly higher than those obtained from the other ICA related features and two other sparse representation features with a small number of training samples via nearest neighbor (NN) classification method.

Attack resistant watermarking technique based on fast curvelet transform and Robust Principal Component Analysis

Watermarking is a technique which embeds the copyright information/identifier to provide authenticity in robust and imperceptible manner. In this paper, a hybrid watermarking technique is proposed based on combination of Fast Curvelet Tansform (FCT), Robust Principal Component Analysis (RPCA) and Singular Value Decomposition (SVD). The gray-scale watermark logo is scrambled using Generalized Arnold Transform (GAT) to enhance the robustness and security. The original image is decomposed to low rank and sparse components using RPCA; the curvelet coefficients are obtained using FCT via Unequally-Spaced Fast Fourier Transforms (USFFT) to embed the processed watermark using SVD into the color image. In curvelet transform, fewer coefficients contain the most energy, also giving optimally sparse representation of the significant image features and edges that helps in efficient recovery of the embedded watermark even after severe image degradation. The robustness and imperceptibility of the proposed technique is verified against a variety of processing operations (noise, filtering) and geometric attacks (crop, resize, projection etc.). The quantitative and visual results reveal that the watermarking technique proposed is more efficient and provides high tolerance against different geometric and image processing attacks.

An Intelligent System Approach for Probabilistic Volume Rendering Using Hierarchical 3D Convolutional Sparse Coding.

In this paper, we propose a novel machine learning-based voxel classification method for highly-accurate volume rendering. Unlike conventional voxel classification methods that incorporate intensity-based features, the proposed method employs dictionary based features learned directly from the input data using hierarchical multi-scale 3D convolutional sparse coding, a novel extension of the state-of-the-art learning-based sparse feature representation method. The proposed approach automatically generates high-dimensional feature vectors in up to 75 dimensions, which are then fed into an intelligent system built on a random forest classifier for accurately classifying voxels from only a handful of selection scribbles made directly on the input data by the user. We apply the probabilistic transfer function to further customize and refine the rendered result. The proposed method is more intuitive to use and more robust to noise in comparison with conventional intensity-based classification methods. We evaluate the proposed method using several synthetic and real-world volume datasets, and demonstrate the methods usability through a user study.

Improving the Impervious Surface Estimation from Hyperspectral Images Using a Spectral-Spatial Feature Sparse Representation and Post-Processing Approach

Impervious surfaces have been widely recognized as an indicator for urbanization and environment monitoring. Plenty of methods have been proposed to extract impervious surfaces using remote sensing images. However, accurately extracting impervious surface is still a challenging task due to the confusion between impervious surface and bare soil. Thus, this paper presents a hybrid approach consisting of spectral-spatial feature sparse representation (SS-SR) and post-processing to extract urban impervious surface from hyperspectral images. We first extracted spectral and spatial features from hyperspectral images. Then, the spectral and spatial information of a pixel is represented by the vector stacking strategy. Each pixel vector can be represented by a linear combination of a few atoms from a learned dictionary, which is more suitable for impervious surface estimation. The sparse coefficients were automatically learned and then used for extracting impervious surface. The proposed impervious surface extraction method was evaluated with four hyperspectral datasets. We compared our algorithms with the state-of-the-art per-pixel based impervious surface extraction methods. The encouraging experimental results demonstrate the SS-SR algorithm generally outperforms the classic support vector machines and random forest. The improvement is more significant when combining SS-SR with post-classification approach.

Sparse-promoting full-waveform inversion based on online orthonormal dictionary learning

Full-waveform inversion (FWI) delivers high-resolution images of the subsurface by minimizing iteratively the misfit between recorded and calculated seismic data. We have attacked this misfit successfully with the Gauss-Newton method and sparsity-promoting regularization based on fixed multiscale transforms that permit significant subsampling of the seismic data when the model perturbation at each FWI data-fitting iteration can be represented with sparse coefficients. Rather than using analytical transforms with predefined dictionaries to achieve sparse representation, we developed an adaptive transform called the sparse orthonormal transform (SOT), whose dictionary is learned from many small training patches taken from the model perturbations in previous iterations. The patch-based dictionary is constrained to be orthonormal and trained with an online approach to provide the best sparse representation of the complex features and variations in the entire model perturbation. The complexity of the training method is proportional to the cube of the number of samples in one small patch. By incorporating compressive subsampling and the adaptive SOT-based representation into the Gauss-Newton least-squares problem for each FWI iteration, the model perturbation can be recovered after an [Formula: see text]-norm sparsity constraint is applied on the SOT coefficients. Numerical experiments on synthetic models determined that the SOT-based sparsity-promoting regularization can provide robust FWI results with reduced computation.

Quantification of 1 H-MRS signals based on sparse metabolite profiles in the time-frequency domain.

MRS is an analytical approach used for both quantitative and qualitative analysis of human body metabolites. The accurate and robust quantification capability of proton MRS (1 H-MRS) enables the accurate estimation of living tissue metabolite concentrations. However, such methods can be efficiently employed for quantification of metabolite concentrations only if the overlapping nature of metabolites, existing static field inhomogeneity and low signal-to-noise ratio (SNR) are taken into consideration. Representation of 1 H-MRS signals in the time-frequency domain enables us to handle the baseline and noise better. This is possible because the MRS signal of each metabolite is sparsely represented, with only a few peaks, in the frequency domain, but still along with specific time-domain features such as distinct decay constant associated with T2 relaxation rate. The baseline, however, has a smooth behavior in the frequency domain. In this study, we proposed a quantification method using continuous wavelet transformation of 1 H-MRS signals in combination with sparse representation of features in the time-frequency domain. Estimation of the sparse representations of MR spectra is performed according to the dictionaries constructed from metabolite profiles. Results on simulated and phantom data show that the proposed method is able to quantify the concentration of metabolites in 1 H-MRS signals with high accuracy and robustness. This is achieved for both low SNR (5dB) and low signal-to-baseline ratio (-5dB) regimes.

An Improved Kernel Minimum Square Error Classification Algorithm Based on $L_{2,1}$ -Norm Regularization

The kernel minimum square error classification (KMSEC) algorithm has been widely used in classification problems. It shows a good performance on image data besides the following drawbacks: not sparse in the solutions and sensitive to noises. The latter drawback will result in a decrease in the recognition performance. To this end, we propose an improved (IKMSEC) by using the $L_{2,1}$ -norm regularization, which can obtain a sparse representation of nonlinear features to guarantee an efficient classification performance. The comprehensive experiments show the promising results in face recognition and image classification.

Double sparse-representation feature selection algorithm for classification

since amount of unlabeled and high-dimensional datasets need to be preprocessed, unsupervised learning plays a more and more important role in machine learning field. This paper proposed a novel un...

Enclosure Transform for Interest Point Detection From Speckle Imagery

We present a fast enclosure transform (ET) to localize complex objects of interest from speckle imagery. This approach explores the spatial confinement on regional features from a sparse image feature representation. Unrelated, broken ridge features surrounding an object are organized collaboratively, giving rise to the enclosureness of the object. Three enclosure likelihood measures are constructed, consisting of the enclosure force, potential energy, and encloser count. In the transform domain, the local maxima manifest the locations of objects of interest, for which only the intrinsic dimension is known a priori. The discrete ET algorithm is computationally efficient, being on the order of O(MN) using N measuring distances across an image of M ridge pixels. It involves easy and few parameter settings. We demonstrate and assess the performance of ET on the automatic detection of the prostate locations from supra-pubic ultrasound images. ET yields superior results in terms of positive detection rate, accuracy and coverage.

Pedestrian detection based on gradient and texture feature integration

In this paper, based on feature integration, we proposed a new method for pedestrian detection. Firstly, we extracted the histogram of oriented gradients (HOG) feature and local binary pattern (LBP) feature from the original images respectively. Secendly, K-singular value decomposition (K-SVD) was used to extract sparse representation features from the HOG and LBP features. Moreover, PCA was used to reduce the dimension of HOG and LBP. Finally, we combined the PCA based features and the K-SVD based sparse representation features directly for fast pedestrian detection in still images. Experimental results on two databases show that the proposed approach is effective for pedestrian detection.

Alzheimer's disease detection via automatic 3D caudate nucleus segmentation using coupled dictionary learning with level set formulation

Background and objectiveThis paper presents a novel method for Alzheimer's disease classification via an automatic 3D caudate nucleus segmentation. MethodsThe proposed method consists of segmentation and classification steps. In the segmentation step, we propose a novel level set cost function. The proposed cost function is constrained by a sparse representation of local image features using a dictionary learning method. We present coupled dictionaries: a feature dictionary of a grayscale brain image and a label dictionary of a caudate nucleus label image. Using online dictionary learning, the coupled dictionaries are learned from the training data. The learned coupled dictionaries are embedded into a level set function. In the classification step, a region-based feature dictionary is built. The region-based feature dictionary is learned from shape features of the caudate nucleus in the training data. The classification is based on the measure of the similarity between the sparse representation of region-based shape features of the segmented caudate in the test image and the region-based feature dictionary. ResultsThe experimental results demonstrate the superiority of our method over the state-of-the-art methods by achieving a high segmentation (91.5%) and classification (92.5%) accuracy. ConclusionsIn this paper, we find that the study of the caudate nucleus atrophy gives an advantage over the study of whole brain structure atrophy to detect Alzheimer's disease.