Video Representation with Dynamic Features from Multi-Frame Frame- Difference Images

이 논문은 고차원의 데이터를 저 차원으로 줄이는 방법 중 하나인 특징추출에 대한 방법들의 특성을 비교한다. 비교대상 방법은 전통적인 PCA(Principal Component Analysis)방법과 시각피질의 특성을 보인다고 알려진 ICA(Independent Component Analysis), 국소기반인식을 구현한 NMF(Non-negative Matrix Factorization), 그리고 이의 성능을 개선한 sNMF(Sparse NMF)로 정하였다. 추출된 특징들의 특성을 시각적으로 확인하기 위하여 필기체 숫자 영상을 대상으로 특징추출을 수행하였으며, 인식기에 적용한 효과의 확인을 위하여 추출된 특징을 다층퍼셉트론에 학습시켜보았다. 각 방법의 특성을 비교한 결과는 응용하고자 하는 문제에서 어떤 특징을 추출하기 원하느냐에 따라 특징추출 방법을 선정할 때 유용할 것이다. In this paper, feature extraction methods, which is one field of reducing dimensions of high-dimensional data, are empirically investigated. We selected the traditional PCA(Principal Component Analysis), ICA(Independent Component Analysis), NMF(Non-negative Matrix Factorization), and sNMF(Sparse NMF) for comparisons. ICA has a similar feature with the simple cell of V1. NMF implemented a "parts-based representation in the brain" and sNMF is a improved version of NMF. In order to visually investigate the extracted features, handwritten digits are handled. Also, the extracted features are used to train multi-layer perceptrons for recognition test. The characteristic of each feature extraction method will be useful when applying feature extraction methods to many real-world problems.

Systematic comparison of unsupervised feature extraction methods for enhancing discrimination accuracy of brown rice flour adulteration in LIBS

Event Abstract Back to Event Comparing independent component analysis and non-negative matrix factorisation in the identification of event-related brain dynamics Sandra Ortega-Martorell1*, Ivan Olier2, Alfredo Vellido3, Paulo J. Lisboa4 and Wael El-Deredy2 1 Universitat Autònoma de Barcelona, Spain 2 University of Manchester, United Kingdom 3 Universitat Politècnica de Catalunya, Spain 4 Liverpool John Moores University, United Kingdom Matrix factorisation methods (such as principal and independent component analysis - PCA, ICA-, or singular value decomposition SVD) are commonly used for decomposing EEG data into factors from which the likely (latent) generating sources could be inferred or identified (Jung et al., 2000; Hoang et al., 2004). Since such factorisations are non-unique, different methods impose different constraints or priors in order to obtain a solution (e.g. components must be orthogonal in PCA, whereas independent in ICA). In non-negative matrix factorisation NMF (Paatero & Tapper,1994; Lee & Seung, 1999), the data matrix, X, is factorised into two non-negative matrices, W and H, whose product provides a good approximation of the original (X ≈ WH). Iterative methods are used to minimise the divergence between X and the product of W and H. In this study, a least squares method (Paatero & Tapper,1994 ) was used. We compared NMF and ICA in terms of: i. Their ability to identify eye movement artifacts in high density EEG electrode arrays. ii. Specificity in the separation of different event related brain dynamics from simultaneously activated sensory modalities. iii. Their computational cost, to assess their suitability for real time processing for the purposes of BCI applications. We used EEG recordings from 15 subjects as in (Brown et al., 2008). NMF was as good as ICA for identifying eye movement and separating different events. The former was much better, though, in the resolution of the components obtained. Computationally, NMF was marginally more efficient than FastICA. Keywords: Brain Signals, EEG Conference: XI International Conference on Cognitive Neuroscience (ICON XI), Palma, Mallorca, Spain, 25 Sep - 29 Sep, 2011. Presentation Type: Poster Presentation Topic: Poster Sessions: Modeling and Analysis of Brain Signals Citation: Ortega-Martorell S, Olier I, Vellido A, Lisboa PJ and El-Deredy W (2011). Comparing independent component analysis and non-negative matrix factorisation in the identification of event-related brain dynamics. Conference Abstract: XI International Conference on Cognitive Neuroscience (ICON XI). doi: 10.3389/conf.fnhum.2011.207.00082 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 16 Nov 2011; Published Online: 25 Nov 2011. * Correspondence: Dr. Sandra Ortega-Martorell, Universitat Autònoma de Barcelona, Barcelona, Spain, S.OrtegaMartorell@ljmu.ac.uk Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Sandra Ortega-Martorell Ivan Olier Alfredo Vellido Paulo J Lisboa Wael El-Deredy Google Sandra Ortega-Martorell Ivan Olier Alfredo Vellido Paulo J Lisboa Wael El-Deredy Google Scholar Sandra Ortega-Martorell Ivan Olier Alfredo Vellido Paulo J Lisboa Wael El-Deredy PubMed Sandra Ortega-Martorell Ivan Olier Alfredo Vellido Paulo J Lisboa Wael El-Deredy Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.

In this paper we explore unsupervised spectral decomposition methods for distinguishing the effect of different spectral components for a set of consecutive spectra from an X-ray binary. We use well-established linear methods for the decomposition, namely principal component analysis, independent component analysis and non-negative matrix factorisation (NMF). Applying these methods to a simulated dataset consisting of a variable multicolour disc black body and a cutoff power law, we find that NMF outperforms the other two methods in distinguishing the spectral components. In addition, due the non-negative nature of NMF, the resulting components may be fitted separately, revealing the evolution of individual parameters. To test the NMF method on a real source, we analyse data from the low-mass X-ray binary GX 339-4 and found the results to match those of previous studies. In addition, we found the inner radius of the accretion disc to be located at the innermost stable circular orbit in the intermediate state right after the outburst peak. This study shows that using unsupervised spectral decomposition methods results in detecting the separate component fluxes down to low flux levels. Also, these methods provide an alternative way of detecting the spectral components without performing actual spectral fitting, which may prove to be practical when dealing with large datasets.

Multivariate prediction of human behavior from resting state data is gaining increasing popularity in the neuroimaging community, with far-reaching translational implications in neurology and psychiatry. However, the high dimensionality of neuroimaging data increases the risk of overfitting, calling for the use of dimensionality reduction methods to build robust predictive models. In this work, we assess the ability of four well-known dimensionality reduction techniques to extract relevant features from resting state functional connectivity matrices of stroke patients, which are then used to build a predictive model of the associated deficits based on cross-validated regularized regression. In particular, we investigated the prediction ability over different neuropsychological scores referring to language, verbal memory, and spatial memory domains. Principal Component Analysis (PCA) and Independent Component Analysis (ICA) were the two best methods at extracting representative features, followed by Dictionary Learning (DL) and Non-Negative Matrix Factorization (NNMF). Consistent with these results, features extracted by PCA and ICA were found to be the best predictors of the neuropsychological scores across all the considered cognitive domains. For each feature extraction method, we also examined the impact of the regularization method, model complexity (in terms of number of features that entered in the model) and quality of the maps that display predictive edges in the resting state networks. We conclude that PCA-based models, especially when combined with L1 (LASSO) regularization, provide optimal balance between prediction accuracy, model complexity, and interpretability.

Since abnormal working conditions or early equipment failures usually cause abnormal performances of critical control loops, loop abnormality must be detected effectively. Motivated by the intrinsic complexity of control loop dynamics, our study introduces a novel scheme that integrates classification and visualization techniques to delve into their dynamic characteristics. The scheme contains correlation analysis, K‐means clustering, and t‐distribution stochastic neighbour embedding (t‐SNE). Firstly, dynamic characteristics from historical data of both manipulated and controlled variables within the control loop were extracted. Subsequently, the correlation analysis method was employed to identify the finite impulse response model. Secondly, various sets of finite impulse response model parameters were compiled into a data matrix for further analysis using the K‐means algorithm to cluster the data effectively. Next, the dimensionality of the labelled data matrix was reduced using t‐SNE for visualization purposes. The scaling process iterated until distinct boundaries emerged for each category, labelling them based on a predefined threshold. Finally, in the online phase, anomalies are diagnosed using the parameters of the finite impulse response model derived from real‐time data, comparing them with the scaled offline model set. The effectiveness of our scheme is validated through the application of real‐world data from a Chinese refinery to identify anomalies within the control loop. Furthermore, our scheme demonstrates superior accuracy when compared with traditional techniques such as principal component analysis (PCA), isometric mapping (ISOMAP), independent component analysis (ICA), and multiscale wavelet transform.

In this paper we propose a new method for removing specular reflection from multispectral dermatological images which is based on blind source separation using Non-negative Matrix Factorization. The key idea of our method is based on a first step which consists to estimate the number of sources involved, by applying a Principal Component Analysis, then a first version of these sources by applying an Independent Component Analysis to our images. We exploit this first version of the sources in a second step to initialize our Non-negative Matrix Factorization algorithm instead of the random initialization that is used by most of the existing methods and that considerably affects their performances. In order to quantify numerically the performance of our method, we also propose a new protocol to artificially mix a specular reflection image with a diffuse reflection image. The tests effected on real and artificial multispectral dermatological images have shown the good performance of our method compared to two of the most used methods.KeywordsMultispectral dermatological imagesRGB imagesSpecular reflectionDiffuse reflectionBlind Source Separation (BSS)Principal Component Analysis (PCA)Independent Component Analysis (ICA)Non-negative Matrix Factorization (NMF)

Nonnegative matrix factorization (NMF) is known as a parts-based linear representation for nonnegative data. This method has been applied for blind source separation (BSS) when the sources are nonnegative. This paper presents a new NMF method for independent component analysis (ICA), which is useful for BSS without the nonnegativity constraint. Using this method, we transform the sources by their cumulative distribution functions (CDFs) and perform the nonparametric quantization to construct a nonnegative matrix where each entry represents the joint probability density of two transformed signals. The NMF procedure is accordingly realized to find the ICA demixing matrix. The independence between sources is maximized towards attaining the uniformity in the joint probability density. In the experiments on the separation of signal and music signals, we show the effectiveness of the proposed NMF-ICA compared to the infomax ICA and FastICA algorithms.

This paper deals with the application of spatially localized, nonoverlapping features for face recognition. The analysis is carried out by using the features generated from two closely related techniques known as independent component analysis (ICA) and nonnegative matrix factorization (NMF). A set of statistically independent basis vectors with sparse features is derived from ICA. Likewise, NMF is used to yield sparse representation of localized features to represent distributed parts over a human face. Similarities between reconstructed faces of test images and a set of synthesised face representations from the basis vectors derived from an image database using the two techniques are measured. The strengths and weaknesses of each method in the context of face recognition are discussed.

Not alike to principal component analysis (PCA) based monitoring statistics (T2 and SPE), the control limits for independent component analysis (ICA) monitoring statistics (I2 , Ie 2 and SPE) cannot be determined directly from a particular approximation distribution due to latent variables do not follow Gaussian distribution. Lee et al. (2004) proposed to use kernel density estimation (KDE) to obtain the control limits. However, the KDE method is very sensitive to the choice of smoothing parameter. Therefore, this study utilizes the support vector machine (SVM) for process fault detection by taking information of ICA extracted statistics as inputs of SVM. The proposed method (named, ICA-SVM) will be implemented in the Tennessee Eastman Process. In which, several multivariate monitoring schemes such as PCA, ICA, modified ICA and ICA- PCA will be also compared to demonstrate the efficiency of proposed method.

Clustering-based initialization for non-negative matrix factorization

The traditional procedure for increasing the signal-tonoise ratio of NMR echo trains is based on stacking and then averaging the data acquired at different depths up to a certain depth interval. Although this works efficiently in increasing the signal-to-noise ratio, it also reduces the vertical resolution of the acquired signal. Therefore a trade-off has to be made between cancelling the noise level without deteriorating the targeted signal but yet still achieving acceptable vertical and spectral resolutions. In NMR logging technology, vertical resolution, usually assessed by NMR-total porosity compared to neutrondensity porosity, is important. However, noise can also affect the spectral distribution of the inverted signal that generates the T1 or T 2 distribution and its components like: clay-bound water, irreducible water, and free fluid. In the area of machine learning applied to denoising, relatively new algorithm developments are using approaches like ICA (Independent Component Analysis), which are based on exploring the statistical independency of the signal components. ICA in fact, puts a stronger mathematical condition on the data-independency—rather than perpendicularity (which traditional Principal Component Analysis (PCA) is based on). However, since ICA is a kind of blind source separation algorithm, the subspace of the ICA solution is usually determined by the relevant numbers of principal components obtained after running PCA. There are several ICA approaches proposed in the literature. One of the most popular is the ‘FastIca’ algorithm. We choose the FastIca algorithm for our test due to its ease of implementation in the MATLAB computational environment. The results from simulated data show that ICA-denoising can be more effective than stacking, if the numbers of the blindly defined ICA components are properly selected. The tradeoff between preserving a significant amount of signal and filtering the right amount of noise is a sensitive aspect in implementing ICA for denoising when rock-fluid characteristics are being explored. Results from real data show that ICA can be readily adapted for post-processing in log analysis, but the selection of ICA-subspace poses a major challenge.

The current dynamic gesture contour feature extraction method has the problems that the recognition rate of dynamic gesture contour feature and the recognition accuracy of dynamic gesture type are low, the recognition time is long, and comprehensive is poor. Therefore, we propose a dynamic gesture contour feature extraction method using residual network transfer learning. Sensors are used to integrate dynamic gesture information. The distance between the dynamic gesture and the acquisition device is detected by transfer learning, the dynamic gesture image is segmented, and the characteristic contour image is initialized. The residual network method is used to accurately identify the contour and texture features of dynamic gestures. Fusion processing weights are used to trace the contour features of dynamic gestures frame by frame, and the contour area of dynamic gestures is processed by gray and binarization to realize the extraction of contour features of dynamic gestures. The results show that the dynamic gesture contour feature recognition rate of the proposed method is 91%, the recognition time is 11.6 s, and the dynamic gesture type recognition accuracy rate is 92%. Therefore, this method can effectively improve the recognition rate and type recognition accuracy of dynamic gesture contour features and shorten the time for dynamic gesture contour feature recognition, and the F value is 0.92, with good comprehensive performance.

Principal components analysis (PCA) has been widely used in many applications, particularly, data compression. Independent component analysis (ICA) has been also developed for blind source separation along with many other applications such as channel equalization, speech processing. Recently, it has been shown that the ICA can be also used for hyperspectral data compression. This paper investigates these two transforms in hyperspectral data compression and further evaluates their strengths and weaknesses in applications of target detection, mixed pixel classification and abundance quantification. In order to take advantage of the strengths of both transform, a new transform, called mixed PCA/ICA transform is developed in this paper. The idea of the proposed mixed PCA/ICA transform is derived from the fact that it can integrate different levels of information captured by the PCA and ICA. In doing so, it combines m principal components (PCs) resulting from the PCA and n independent components (ICs) generated by the ICA to form a new set of (m+n) mixed components used for hyperspectral data compression. The resulting transform is referred to as mixed (m,n)-PCA/ICA transform. In order to determine the total number of components, p needed to be generated for the mixed (m,n)-PCA/ICA transform, a recently developed virtual dimensionality (VD) is introduced to estimate the p where p = m + n. If m = p and n = 0, then mixed (m,n)-PCA/ICA transform is reduced to PCA transform. On the other hand, if m = 0 and n = p, then mixed (m,n)-PCA/ICA transform is reduced to ICA. Since various combinations of m and n have different impacts on the performance of the mixed PCA/ICA spectral/spatial compression in applications, experiments based on subpixel detection and mixed pixel quantification are conducted for performance evaluation.

<p indent=0mm>The existing sketch-based 3D model retrieval methods often regard data as static input, and utilize convolutional neural network to extract features, which ignore the dynamic attributes of input data, resulting in partial loss of useful information and not ideal retrieval effect. In order to solve this problem, a spatiotemporal information joint embedding-based end-to-end sketch-3D model retrieval method is proposed. Firstly, the sketch is represented as a dynamic drawing sequence, which reflects the temporal information contained in the drawing process; meanwhile, the 3D model is represented as a multi-view sequence to reflect the position relationship between views. Secondly, an end-to-end dual-stream network including static spatial feature extraction and dynamic time series feature extraction is constructed. Combined with triplet central metric learning, a joint spatiotemporal feature embedding of cross domain data is established to fully capture the static and dynamic features contained in sketches and 3D models, and reduce the difference between cross domain. Finally, experiments are carried out on the standard public data sets SHREC2013 and SHREC2014. Compared with the existing work, the accuracy rate is higher, which verifies the feasibility and effectiveness of the proposed algorithm.