Image Representation Method Research Articles

Artificial Intelligence Recognition Simulation of 3D Multimedia Visual Image Based on Sparse Representation Algorithm

With the rapid development of computer networks and multimedia technologies, images, which are important carriers of information dissemination, have made human cognition of things easier. Image recognition is a basic research task in computer vision, multimedia search, image understanding and other fields. This paper proposes a hierarchical feature learning structure that is completely automatically based on the original pixels of the image, and uses the K-SVD (K-Singular Value Decomposition) algorithm with label consistency constraints to train the discriminant dictionary. For different types of image data sets, the algorithm only extracts image blocks. After dense sampling, an efficient OMP (Orthogonal Matching Pursuit) encoder is used to obtain a layered sparse representation. The improved SIFT (Scale Invariant Feature Transform) algorithm is used to solve the difficult problem of multimedia visual image stereo matching. The feature point extraction and stereo matching of multimedia visual images, different scales and different viewpoint images are analyzed separately. Aiming at a large number of low-dimensional geometric features of 3D images, this paper studies the extraction and sorting strategies of low-dimensional geometric features of 3D images. A sparse representation method for 3D images is proposed, and the sparseness of image features is evaluated. This further improves the accuracy of 3D image representation and the robustness of 3D image recognition algorithms.

Intelligent Diagnosis for Railway Wheel Flat Using Frequency-Domain Gramian Angular Field and Transfer Learning Network

The intelligent diagnosis of wheel flat based on vibration image classification is a promising research subject for performance maintenance of railway vehicles. However, the image representation method of vibration signal and classification network construction under small samples have become two obstacles to intelligent diagnosis of wheel flat. This paper presents a novel frequency-domain Gramian angular field (FDGAF) algorithm to encode the vibration signal of wheel flat to featured images. Furthermore, a modified transfer learning network is introduced to classify these featured images under small samples without any involvement of prior knowledge. The proposed FDGAF can calculate the Gramian angular matrix of axle box acceleration signal in frequency domain and assign frequency position dependence to the featured images to preserve original characteristic information. Then, these featured images can be intelligent classified by a transfer learning network under the condition of 30 sample without require of prior knowledge. To verify the efficiency of this proposed method, 12 cases of artificial wheel flats are processed on a scaled railway test rig, and their axle box acceleration signals are collected to obtain visual diagnosis results. The verfication proves that FDGAF is able to obtain accurate diagnostic results with high separability, for separability indexes of FDGAF reaches 10.8, 8.7, 14.9, and 5.8. We anticipate that this method will find use in the performance maintenance of railway vehicles and the improvement of industrial condition monitoring.

A novel numerical representation for proteins: Three-dimensional Chaos Game Representation and its Extended Natural Vector

Chaos Game Representation (CGR) was first proposed to be an image representation method of DNA and have been extended to the case of other biological macromolecules. Compared with the CGR images of DNA, where DNA sequences are converted into a series of points in the unit square, the existing CGR images of protein are not so elegant in geometry and the implications of the distribution of points in the CGR image are not so obvious. In this study, by naturally distributing the twenty amino acids on the vertices of a regular dodecahedron, we introduce a novel three-dimensional image representation of protein sequences with CGR method. We also associate each CGR image with a vector in high dimensional Euclidean space, called the extended natural vector (ENV), in order to analyze the information contained in the CGR images. Based on the results of protein classification and phylogenetic analysis, our method could serve as a precise method to discover biological relationships between proteins.

A new weighted two-dimensional vector quantisation encoding method in bag-of-features for histopathological image classification

Automated histopathological image analysis is a challenging problem due to the complex morphological structure of histopathology images. Bag-of-features is one of the prominent image representation methods which has been successfully applied in histopathological image analysis. There are four phases in the bag-of-features method, namely feature extraction, codebook construction, feature encoding, and classification. Out of which feature encoding is one of the prime phases. In feature encoding phase, images are represented in terms of visual words before feeding into support vector machine classifier. However, the feature encoding phase of the bag-of-features framework considers the one feature to encode each image in terms of visual words due to which the system can not use the merits of other features. Therefore, to improve the efficacy of the bag-of-features framework, a new weighted two-dimensional vector quantisation encoding method is proposed in this work. The proposed method is tested on two histopathological image datasets for classification. The experimental results show that the combination of SIFT and ORB features with two dimensional vector quantisation encoding method returns 80.13% and 77.13% accuracy on ADL and Blue Histology datasets respectively which is better than other considered encoding methods.

A Novel Circular Flexible Multiscale Geometric Analysis Method

Image representation is an essential problem in image processing. The most effective image representation method is multiscale geometric analysis (MGA). However, the current representative MGA method has some shortcomings, such as the fixed division of the scale spectrum and the direction spectrum, which cannot achieve well the sparsest representation of the image; the lack of rotation invariance and translation invariance make it impossible to achieve better results in the application. In this paper, a new circular flexible MGA method is proposed to solve this problem. First, the McClellan method was used to design a circular flexible multiscale decomposition, which had translation invariance and rotation invariance due to the circular spectral shape. Moreover, the spectral division of the new multiscale decomposition method could be flexibly changed according to the image, rather than being fixed at π/2 in the traditional manner. Second, new linear phase directional filter banks were proposed to divide the direction flexibly, which had selectivity in any direction. Finally, the two steps above were combined to construct a new MGA method, which maintained translation invariance, low redundancy, and flexibility in both scale and direction, resulting in better frequency localization and more sparse representation. The results of our denoising experiment show that the method proposed in this paper not only achieves a perfect reconstruction, but also performs well in image denoising compared to other state-of-the-art MGA methods.

Deep Non-Negative Matrix Factorization Architecture Based on Underlying Basis Images Learning

The non-negative matrix factorization (NMF) algorithm represents the original image as a linear combination of a set of basis images. This image representation method is in line with the idea of "parts constitute a whole" in human thinking. The existing deep NMF performs deep factorization on the coefficient matrix. In these methods, the basis images used to represent the original image is essentially obtained by factorizing the original images once. To extract features reflecting the deep localization characteristics of images, a novel deep NMF architecture based on underlying basis images learning is proposed for the first time. The architecture learns the underlying basis images by deep factorization on the basis images matrix. The deep factorization architecture proposed in this paper has strong interpretability. To implement this architecture, this paper proposes a deep non-negative basis matrix factorization algorithm to obtain the underlying basis images. Then, the objective function is established with an added regularization term, which directly constrains the basis images matrix to obtain the basis images with good local characteristics, and a regularized deep non-negative basis matrix factorization algorithm is proposed. The regularized deep nonlinear non-negative basis matrix factorization algorithm is also proposed to handle pattern recognition tasks with complex data. This paper also theoretically proves the convergence of the algorithm. Finally, the experimental results show that the deep NMF architecture based on the underlying basis images learning proposed in this paper can obtain better recognition performance than the other state-of-the-art methods.

Quantum Representation of Indexed Images and its Applications

Indexed images have been widely used in classical digital image processing, however there have been no reports on the representation and applications of indexed images in quantum image processing so far. To solve the representation problem of indexed images on a quantum computer, a quantum indexed image representation (QIIR) method is proposed in the paper. A quantum indexed image consists of a quantum data matrix and a quantum palette matrix. Each data structure is based on the basic states of qubit sequence to represent information, including pixel positions and pixel values in the data matrix, and indexes and color values in the palette matrix. Several simple geometric and color transformations of quantum indexed images are presented later, including orthogonal rotation, cyclic shift, color inversion, color replacement and color look-up. Finally, a quantum indexed image steganography based on EzStego is proposed. In this scheme, the distance between two arbitrary color values in the quantum palette is first calculated, and then several effective color pairs are obtained. At last, according to embedded message bits, pixel values in the data matrix are updated in light of effective color pairs, and a new quantum data matrix with embedded message is obtained. The proposed scheme can be executed on a future quantum computer. The feasibility and validness of this scheme are verified on a classical computer from four aspects: visual quality, embedding capacity, robustness and security.

Maximum Correntropy Criterion-Based Robust Semisupervised Concept Factorization for Image Representation.

Concept factorization (CF) has shown its great advantage for both clustering and data representation and is particularly useful for image representation. Compared with nonnegative matrix factorization (NMF), CF can be applied to data containing negative values. However, the performance of CF method and its extensions will degenerate a lot due to the negative effects of outliers, and CF is an unsupervised method that cannot incorporate label information. In this article, we propose a novel CF method, with a novel model built based on the maximum correntropy criterion (MCC). In order to capture the local geometry information of data, our method integrates the robust adaptive embedding and CF into a unified framework. The label information is utilized in the adaptive learning process. Furthermore, an iterative strategy based on the accelerated block coordinate update is proposed. The convergence property of the proposed method is analyzed to ensure that the algorithm converges to a reliable solution. The experimental results on four real-world image data sets show that the new method can almost always filter out the negative effects of the outliers and outperform several state-of-the-art image representation methods.

A Node2Vec-Based Gene Expression Image Representation Method for Effectively Predicting Cancer Prognosis

A Node2Vec-Based Gene Expression Image Representation Method for Effectively Predicting Cancer Prognosis

Multi‐scale microstructure binary pattern extraction and learning for image representation

In this study, an image representation method based on multi-scale microstructural binary pattern extraction is proposed, which uses zero-mean microstructural pattern binarisation. This method can express all kinds of important pattern structures that may appear in the image. By using the dominant binary pattern learning model, the dominant feature pattern sets adapted to different datasets can be obtained, which have good performance in the aspects of feature robustness, recognition, and representation ability. This method can greatly reduce the dimension of feature coding and improve the speed of the algorithm. The experimental results show that this method has strong recognition ability and robustness, is superior to the traditional local binary pattern and grey image micorstructure maximum response pattern methods, and has a competitive performance compared with the results of many latest algorithms.

A novel bearing fault diagnosis method based on 2D image representation and transfer learning-convolutional neural network

Traditional methods used for intelligent condition monitoring and diagnosis significantly depend on manual feature extraction and selection. To address this issue, a transfer learning-convolutional neural network (TLCNN) based on AlexNet is proposed for bearing fault diagnosis. Firstly, a 2D image representation method converts vibration signals to 2D time-frequency images. Secondly, the proposed TLCNN model extracts the features of the 2D time-frequency images and achieves the classification conditions of the bearing, which is faster to train and more accurate. Thirdly, t-distributed stochastic neighbor embedding (t-SNE) is applied to visualize the feature learning process to demonstrate the feature learning ability of the proposed model. The experimental results verify that the proposed fault diagnosis model has higher accuracy and has much better robustness against noise than other deep learning and traditional methods.

Supervised Deep Feature Embedding With Handcrafted Feature.

Image representation methods based on deep convolutional neural networks (CNNs) have achieved the state-of-the-art performance in various computer vision tasks, such as image retrieval and person re-identification. We recognize that more discriminative feature embeddings can be learned with supervised deep metric learning and handcrafted features for image retrieval and similar applications. In this paper, we propose a new supervised deep feature embedding with a handcrafted feature model. To fuse handcrafted feature information into CNNs and realize feature embeddings, a general fusion unit is proposed (called Fusion-Net). We also define a network loss function with image label information to realize supervised deep metric learning. Our extensive experimental results on the Stanford online products' data set and the in-shop clothes retrieval data set demonstrate that our proposed methods outperform the existing state-of-the-art methods of image retrieval by a large margin. Moreover, we also explore the applications of the proposed methods in person re-identification and vehicle re-identification; the experimental results demonstrate both the effectiveness and efficiency of the proposed methods.

Selective multi-descriptor fusion for face identification

Over the last 2 decades, face identification has been an active field of research in computer vision. As an important class of image representation methods for face identification, fused descriptor-based methods are known to lack sufficient discriminant information, especially when compared with deep learning-based methods. This paper presents a new face representation method, multi-descriptor fusion (MDF), which represents face images through a combination of multiple descriptors, resulting in hyper-high dimensional fused descriptor features. MDF enables excellent performance in face identification, exceeding the state-of-the-art, but it comes with high memory and computational costs. As a solution to the high cost problem, this paper also presents an optimisation method, discriminant ability-based multi-descriptor selection (DAMS), to select a subset of descriptors from the set of 65 initial descriptors whilst maximising the discriminant ability. The MDF face representation, after being refined by DAMS, is named selective multi-descriptor fusion (SMDF). Compared with MDF, SMDF has much smaller feature dimension and is thus usable on an ordinary PC, but still has similar performance. Various experiments are conducted on the CAS-PEAL-R1 and LFW datasets to demonstrate the performance of the proposed methods.

Quaternion-based weighted nuclear norm minimization for color image denoising

The quaternion method plays an important role in color image processing, because it represents the color image as a whole rather than as a separate color space component, thus naturally handling the coupling among color channels. The weighted nuclear norm minimization (WNNM) scheme assigns different weights to different singular values, leading to more reasonable image representation method. In this paper, we propose a novel quaternion weighted nuclear norm minimization (QWNNM) model and algorithm under the low rank sparse framework. The proposed model represents the color image as a low rank quaternion matrix, where quaternion singular value decomposition can be calculated by its equivalent complex matrix. We solve the QWNNM by adaptively assigning different singular values with different weights. Color image denoising is implemented by QWNNM based on non-local similarity priors. In this new color space, the inherent color structure can be well preserved during image reconstruction. For high noise levels, we apply a Gaussian low pass filter (LPF) to the noisy image as a preprocessing before QWNNM, which reduces the iteration numbers and improves the denoised results. The experimental results clearly show that the proposed method outperforms K-SVD, QKSVD and WNNM in terms of both quantitative criteria and visual perceptual.

Learning binary hash codes for finger vein image retrieval

High dimensional real-valued features have been shown to be effective for finger vein identification, but result in high computational cost especially in a large-scale finger vein database. Therefore compact binary codes are generally used to obtain fast query speed as well as reduce storage requirements. However, only a few strategies are available for finger vein retrieval in Hamming space, and existing binary representation of finger veins can become unstable and undiscriminating due to finger vein variations induced by translation, rotation and illumination. In this paper, we propose a binary hash codes learning algorithm to map finger vein images in the original feature space to Hamming space. First, to obtain the discriminative finger vein image features, a novel finger vein image representation method called Nonlinearly Subspace Coding (NSC) is proposed. The codebook is a union of low-dimensional linear subspaces instead of visual words. For a given local finger vein texton, its top-k nearest subspaces are found and the texton is nonlinearly mapped into these subspaces. Then, a finger vein Binary Hash Codes (BHC) learning method is proposed by jointly considering the discriminability and the stability of the binary code. Experimental results on a two-session public finger vein database and a large fused finger vein database demonstrate the effectiveness and efficiency of our binary hash coding learning algorithm for large-scale finger vein retrieval.

A novel attention-based hybrid CNN-RNN architecture for sEMG-based gesture recognition.

The surface electromyography (sEMG)-based gesture recognition with deep learning approach plays an increasingly important role in human-computer interaction. Existing deep learning architectures are mainly based on Convolutional Neural Network (CNN) architecture which captures spatial information of electromyogram signal. Motivated by the sequential nature of electromyogram signal, we propose an attention-based hybrid CNN and RNN (CNN-RNN) architecture to better capture temporal properties of electromyogram signal for gesture recognition problem. Moreover, we present a new sEMG image representation method based on a traditional feature vector which enables deep learning architectures to extract implicit correlations between different channels for sparse multi-channel electromyogram signal. Extensive experiments on five sEMG benchmark databases show that the proposed method outperforms all reported state-of-the-art methods on both sparse multi-channel and high-density sEMG databases. To compare with the existing works, we set the window length to 200ms for NinaProDB1 and NinaProDB2, and 150ms for BioPatRec sub-database, CapgMyo sub-database, and csl-hdemg databases. The recognition accuracies of the aforementioned benchmark databases are 87.0%, 82.2%, 94.1%, 99.7% and 94.5%, which are 9.2%, 3.5%, 1.2%, 0.2% and 5.2% higher than the state-of-the-art performance, respectively.

Unsupervised feature learning via prior information‐based similarity metric learning for face verification

Here, an efficient framework is developed to address the problem of unconstrained face verification. In particular, an unsupervised feature learning method for face image representation and a novel similarity metric model are discussed. First, the authors propose an unsupervised feature learning method with sparse auto-encoder (SAE) based on local descriptor (SAELD). A set of filter operators are learned based on SAE model from local patches, and face descriptors are extracted by applying the set of filter operators to convolve images. This can address the face discriminative representation issue of unconstrained face verification. Then pairwise SAELD descriptors are projected into the weighted subspace. Furthermore, a prior information-based similarity metric learning model is presented, in which the metric matrix is learned by enforcing a regularisation term based on the prior similar and discriminative information. This idea can improve the robustness to intra-personal variations and discrimination to inter-personal variations. Experimental results show that the proposed method has competitive performance compared with several state-of-the-art methods on challenging labelled faces in the wild data set.

Iterative Weighted Non-smooth Non-negative Matrix Factorization for Face Recognition

Non-negative Matrix Factorization (NMF) is a part-based image representation method. It comes from the intuitive idea that entire face image can be constructed by combining several parts. In this paper, we propose a framework for face recognition by finding localized, part-based representations, denoted “Iterative weighted non-smooth non-negative matrix factorization” (IWNS-NMF). A new cost function is proposed in order to incorporate sparsity which is controlled by a specific parameter and weights of feature coefficients. This method extracts highly localized patterns, which generally improves the capability of face recognition. After extracting patterns by IWNS-NMF, we use principle component analysis to reduce dimension for classification by linear SVM. The Recognition rates on ORL, YALE and JAFFE datasets were 97.5, 93.33 and 87.8%, respectively. Comparisons to the related methods in the literature indicate that the proposed IWNS-NMF method achieves higher face recognition performance than NMF, NS-NMF, Local NMF and SNMF.

Bundled Local Features for Image Representation

Local features have been widely used for image representation. Traditional methods often treat each local feature independently or simply model the correlations of local features with spatial partition. However, local features are correlated and should be jointly modeled. Besides, due to the variety of images, predefined partition rules will probably introduce noisy information. To solve these problems, in this paper we propose a novel bundled local features method for efficient image representation and apply it for classification. Specially, we first extract local features and bundle them together with over-complete spatial shapes by viewing each local feature as the central point. Then, the most discriminatively bundling features are selected by reconstruction error minimization. The encoding parameters are then used for image representations in a matrix form. Finally, we train bi-linear classifiers with quadratic hinge loss to predict the classes of images. The proposed method can combine local features appropriately and efficiently for discriminative representations. Experimental results on three image data sets show the effectiveness of the proposed method compared with other local features combination strategies.

Efficient feature-based image registration by mapping sparsified surfaces

With the advancement in the digital camera technology, the use of high resolution images and videos has been widespread in the modern society. In particular, image and video frame registration is frequently applied in computer graphics and film production. However, conventional registration approaches usually require long computational time for high resolution images and video frames. This hinders the application of the registration approaches in the modern industries. In this work, we first propose a new image representation method to accelerate the registration process by triangulating the images effectively. For each high resolution image or video frame, we compute an optimal coarse triangulation which captures the important features of the image. Then, we apply a surface registration algorithm to obtain a registration map which is used to compute the registration of the high resolution image. Experimental results suggest that our overall algorithm is efficient and capable to achieve a high compression rate while the accuracy of the registration is well retained when compared with the conventional grid-based approach. Also, the computational time of the registration is significantly reduced using our triangulation-based approach.