Добірка наукової літератури з теми "Multiway CCA"

Canonical correlation analysis (CCA) has been used for the steady-state visual evoked potential (SSVEP) based brain-computer interface (BCI) for a long time. However, the reference signal of CCA is relatively simple and lacks subject-specific information. Moreover, over-fitting may occur when a short time window (TW) length was used in CCA. In this article, an optimized L1-regularized multiway canonical correlation analysis (L1-MCCA) is combined with a support vector machine (SVM) to overcome the aforementioned shortcomings in CCA. The correlation coefficients obtained by L1-MCCA were transferred into a particle-swarm-optimization (PSO)-optimized support vector machine (SVM) classifier to improve the classification accuracy. The performance of the proposed method was evaluated and compared with the traditional CCA and power spectral density (PSD) methods. The results showed that the accuracy of the L1-MCCA-PSO-SVM was 96.36% and 98.18% respectively when the TW lengths were 2 s and 6 s. This accuracy is higher than that of the traditional CCA and PSD methods.

Canonical correlation analysis (CCA) has been one of the most popular methods for frequency recognition in steady-state visual evoked potential (SSVEP)-based brain–computer interfaces (BCIs). Despite its efficiency, a potential problem is that using pre-constructed sine-cosine waves as the required reference signals in the CCA method often does not result in the optimal recognition accuracy due to their lack of features from the real electro-encephalo-gram (EEG) data. To address this problem, this study proposes a novel method based on multiset canonical correlation analysis (MsetCCA) to optimize the reference signals used in the CCA method for SSVEP frequency recognition. The MsetCCA method learns multiple linear transforms that implement joint spatial filtering to maximize the overall correlation among canonical variates, and hence extracts SSVEP common features from multiple sets of EEG data recorded at the same stimulus frequency. The optimized reference signals are formed by combination of the common features and completely based on training data. Experimental study with EEG data from 10 healthy subjects demonstrates that the MsetCCA method improves the recognition accuracy of SSVEP frequency in comparison with the CCA method and other two competing methods (multiway CCA (MwayCCA) and phase constrained CCA (PCCA)), especially for a small number of channels and a short time window length. The superiority indicates that the proposed MsetCCA method is a new promising candidate for frequency recognition in SSVEP-based BCIs.

In this study, the statistical method of the Canonical Correlation Analysis (CCA) was applied in order to gain insight into the interaction of nearshore wave climate and the morphological evolution of the seabed at Lubiatowo beach in Poland. Lubiatowo, a non-tidal, sandy beach characterized by multiple longshore bars (Figure 1) is typical of Southern Baltic coasts. In the past, this method has also been used in the coastal field by Larson et al. (2000). Rozinsky (2003) also used the CCA to find a correlation between components of a multibar system and the interaction between the multiple longshore bars. Elsewhere, Horrillo-Caraballo and Reeve (2008), examined the performance of CCA-based at Duck (USA).

The perception of speech and audio is one of the defining features of humans. Much of the brain’s underlying processes as we listen to acoustic signals are unknown, and significant research efforts are needed to unravel them. The non-invasive recordings capturing the brain activations like electroencephalogram (EEG) and magnetoencephalogram (MEG) are commonly deployed to capture the brain responses to auditory stimuli. But these non-invasive techniques capture artifacts and signals not related to the stimuli, which distort the stimulus-response analysis. The effect of the artifacts be- comes more evident for naturalistic stimuli. To reduce the inter-subject redundancies and amplify the components related to the stimuli, the EEG responses from multiple subjects listening to a common naturalistic stimulus need to be normalized. The currently used normalization and pre-processing methods are the canonical correlation analysis (CCA) models and the temporal response function based forward/backward models. However, these methods assume a simplistic linear relationship between the audio features and the EEG responses and therefore, may not alleviate the recording artifacts and interfering signals in EEG. This thesis proposes novel methods using machine learning advances to improve the audio-EEG analysis. We propose a deep learning framework for audio-EEG analysis in intra-subject and inter-subject settings. The deep learning based intra-subject analysis methods are trained with a Pearson correlation-based cost function between the stimuli and EEG responses. This model allows the transformation of the audio and EEG features that are maximally correlated. The correlation-based cost function can be optimized with the learnable parameters of the model trained using standard gradient descent- based methods. This model is referred to as the deep CCA (DCCA) model. Several experiments are performed on the EEG data recorded when the subjects are listening to naturalistic speech and music stimuli. We show that the deep methods obtain better representations than the linear methods and results in statistically significant improvements in correlation values. Further, we propose a neural network model with shared encoders that align the EEG responses from multiple subjects listening to the same audio stimuli. This inter-subject model boosts the signals common across the subjects and suppresses the subject-specific artifacts. The impact of improving stimulus-response correlations are highlighted based on multi-subject EEG data from speech and music tasks. This model is referred to as the deep multi-way canonical correlation analysis (DMCCA). The combination of inter-subject analysis using DMCCA and intra-subject analysis using DCCA is shown to provide the best stimulus-response in audio-EEG experiments. We highlight how much of the audio signal can be recovered purely from the non- invasive EEG recordings with modern machine learning methods, and conclude with a discussion on future challenges in audio-EEG analysis.

Poly (methyl methacrylate) (PMMA) and acrylonitrile-butadiene-styrene (ABS) – multiwall carbon nanotube (MWNT) and chopped carbon fiber (CCF) composites were prepared by a melt mixing protocol at various concentrations. Specimens were fabricated and tested using constant amplitude-of-deflection fatigue testing. The numbers of cycles to failure were recorded and analyzed using the linear version of the 2-parameter Weibull model. In the PMMA matrix, the 1.0vol% MWNT reinforced composites outperformed the neat PMMA matrix by +396% while the 1.0vol% CCF composites increased fatigue life by +198% over the control. The increase in fatigue life may be attributed to the nanoscale dimensions of the MWNTs. This enables them to directly interact with the matrix at the sub-micron scale where damage such as crazing begins, which ultimately initiates a critical crack that leads to failure of the specimen. The ABS composite specimens did not show any increase in fatigue life. The underlying reasons for the lack of fatigue improvement remain unclear.

Using 30-psec pulses from a frequency doubled Nd:YAG laser, we have amplified and detected images with a resolution of less than 300 μm through suspensions of latex spheres and non-dairy creamer with nonscattered extinction ratios of up to e - 40. Our time-gated image amplifier can produce images under conditions in which the scattering medium is sufficiently dense that an image cannot be seen on either a streak camera or through a long-time integration on a CCD camera. The 532-nm frequency-doubled output was used both to generate a seed-Stokes pulse at 683 nm using hydrogen at 30 atm and to pump the 1-m long Raman amplifier.1 The seed-Stokes light was used to illuminate a bar chart of line spacing 0.315 mm placed immediately before a cell containing the scattering medium. The direct image, or its Fourier transform, was amplified in the Raman amplifier. The original bar chart image was then observed using an intensified camera. By using a delay to vary the time of arrival of the pump pulse at the amplifier we were able to selectively amplify the image-bearing component at the front of the pulse and so discriminate against the multiply scattered light which can extend out for many pulse durations.

Abstract Global warming is a critical issue that has garnered significant attention during the last few decades. This temperature increase results from the emission of greenhouse gases into the atmosphere, most notably CO2. Carbon Capture and Storage (CCS) technology has been established as one of the most successful ways to store CO2 in underground layers and prevent it from being released into the atmosphere. However, CO2 is difficult to contain in subterranean layers subjected to high-pressure, high temperature (HPHT) conditions due to the degradation of Portland cement caused by chemical interaction with wet/dry CO2. Numerous studies have been conducted to increase the cement's resistance to CO2 attack, but limited effectiveness has been found when these methods have been evaluated under various settings. Given the distinctive properties of nanomaterials, such as high surface areas, quick contact, and resilience to heat, Nano Glass Flake (NGF) and Multiwall Carbon Nano Tube (MWCNT) were deemed to be suitable additional materials for increasing cement efficiency. On cement samples treated with NGFs and MWCNTs, a number of pre-carbonation and post-carbonation tests were performed. The pre-carbonation tests revealed that the density of NGFs-based cement remained constant with that of neat cement while the plastic viscosity increased. Additionally, it was recommended not to add more than 1wt% NGFs to the cement, as this would result in a high viscosity paste, which would negatively affect the pumping operation. On the other hand, this threshold for the viscosity of MWCNTs was roughly 0.25wt%. It was found that by using nanoparticles and employing a proper dispersion process, the cement's overall physical performance can be improved, and a lower amount of Portlandite is formed, which is critical for increased resistance to CO2 attack. In a static reactor, samples with the best pre-carbonation performance were subjected to water saturated supercritical CO2 for 56 days. It was then discovered that CO2 diffuses into cement and increases cement decomposition in the post-carbonation stage of the experiment. Samples weighing more than 0.5wt %. NGFs and 0.05 wt% MWCNTs had the smallest carbonated regions, indicating carbonated cement. However, the number of nanoparticles added to each sample resulted in a variable level of carbonation. Cement made using MWCNTs has a higher compressive strength due to its ability to manipulate CaCO3 crystal shape. NGFs-based cement, on the other hand, could be a better solution in terms of CO2 resistance. Due to their substantially lower cost than MWCNTs, it is possible to increase cement performance in CCS operations without imposing a high cost on projects.

There are certain applications where it is desirable to image an object embedded in, or hidden behind, a scattering medium. An example in the medical field would be transillumination of human tissue1,2. Among other potential applications one could consider imaging through any highly scattering medium like clouds or plastics. The obstacle to overcome is the multiply scattered light obscuring the image of the object. Several techniques have been employed to solve this problem. All of them are relying on the fact that the highly scattered light is delayed in time with respect to the less scattered light. The forward scattered component of the light transmits more spatial information about the hidden object. Using very short laser pulses(typically less than a few ps) and time gating one can filter out the scattered light which arrives at the detector later in time. Various time-gating schemes have been used including fast gated electronic amplifier2, fast streak camera3, optical shutters using Kerr effect4, holography with either short5 or long6 broadband laser pulses, and a number of nonlinear-optical techniques7,8,9. Often submillimeter spatial resolution is desired which requires the use of subpicosecond lasers with the gating techniques. It is desirable to extend the hidden object imaging to longer laser pulses that are readily available with commercial lasers today. For example, with longer pulses, higher energies per pulse can be used within the ANSI exposure limits for human tissue. This can significantly reduce the requirements on image collection. Recently a technique called Chrono-Coherent Imaging6 (CCI) was developed, where the short coherence length of a long pulse duration broadband light source is used for gating. The light is split in two beams, where one of the beams illuminates the object. The interference between the two beams is recorded on the film as a holographic image. One of the disadvantages of this method is the degradation of the holographic image due to the scattered light which is also incident on the film. In addition, it is not a real-time imaging technique.