This letter describes a fast and very simple algorithm for estimating the fetal electrocardiogram (FECG). It is based on independent component analysis, but we substitute its computationally demanding calculations for a much simpler procedure. The resulting method consists of two steps: 1) a dimensionality reduction step and 2) a computationally light postprocessing stage used to enhance the FECG signal.
This letter presents a successful system for recovering the fetal electrocardiogram using multidimensional ICA (MICA). MICA requires as many observations as sources. To increase the number of observations, MICA is often applied to data sets that include measurements taken at the mother's thoracic region. However, experiments suggest that the propagation from the maternal heart to the mother's abdomen is not a simple delay, and that approach may fail. Alternatively, our method first estimates the maternal ECG directly from the mother's abdomen. Then, inputs this estimated ECG to MICA. Experiments show superior performance as compared with the traditional approach.
22 June 2016 in IEEE Transactions on Pattern Analysis and Machine Intelligence
Principal component analysis (PCA) based on L1-norm maximization is an emerging technique that has drawn growing interest in the signal processing and machine learning research communities, especially due to its robustness to outliers. The present work proves that L1-norm PCA can perform independent component analysis (ICA) under the whitening assumption. However, when the source probability distributions fulfil certain conditions, the L1-norm criterion needs to be minimized rather than maximized, which can be accomplished by simple modifications on existing optimal algorithms for L1-PCA. If the sources have symmetric distributions, we show in addition that L1-PCA is linked to kurtosis optimization. A number of numerical experiments illustrate the theoretical results and analyze the comparative performance of different algorithms for ICA via L1-PCA. Although our analysis is asymptotic in the sample size, this equivalence opens interesting new perspectives for performing ICA using optimal algorithms for L1-PCA with guaranteed global convergence while inheriting the increased robustness to outliers of the L1-norm criterion.
11 April 2019 in IEEE Transactions on Neural Systems and Rehabilitation Engineering
In brain-computer interfaces the typical models of the EEG observations usually lead to a poor estimation of the trial covariance matrices, given the high non-stationarity of the EEG sources. We propose the application of two techniques that significantly improve the accuracy of these estimations and can be combined with a wide range of motor imagery BCI methods. The first one scales the observations in such a way that implicitly normalizes the common temporal strength of the source activities. When the scaling applies independently to the trials of the observations the procedure justifies and improves the classical preprocessing for the EEG data. Additionally, when the scaling is instantaneous and independent for each sample, the procedure particularizes to Tyler’s method in statistics for obtaining a distribution-free estimate of scattering. In this case, the proposal provides an original interpretation of this existing method as a technique that pursuits an implicit instantaneous power-normalization of the underlying source processes. The second technique applies to the classifier and improves its performance through a convenient regularization of the features covariance matrix. Experimental tests reveal that a combination of the proposed techniques with state-of-the-art algorithms for motor-imagery classification provides a significant improvement in the classification results.
Brain computer interfaces (BCIs) have been attracting a great interest in recent years. The common spatial patterns (CSP) technique is a well-established approach to the spatial filtering of the electroencephalogram (EEG) data in BCI applications. Even though CSP was originally proposed from a heuristic viewpoint, it can be also built on very strong foundations using information theory. This paper reviews the relationship between CSP and several information-theoretic approaches, including the Kullback–Leibler divergence, the Beta divergence and the Alpha-Beta log-det (AB-LD)divergence. We also revise other approaches based on the idea of selecting those features that are maximally informative about the class labels. The performance of all the methods will be also compared via experiments.
This paper presents a system with location functionalities for the inventory of traffic signs based on passive RFID technology. The proposed system simplifies the current video-based techniques, whose requirements regarding visibility are difficult to meet in some scenarios, such as dense urban areas. In addition, the system can be easily extended to consider any other street facilities, such as dumpsters or traffic lights. Furthermore, the system can perform the inventory process at night and at a vehicle’s usual speed, thus avoiding interfering with the normal traffic flow of the road. Moreover, the proposed system exploits the benefits of the passive RFID technologies over active RFID, which are typically employed on inventory and vehicular routing applications. Since the performance of passive RFID is not obvious for the required distance ranges on these in-motion scenarios, this paper, as its main contribution, addresses the problem in two different ways, on the one hand theoretically, presenting a radio wave propagation model at theoretical and simulation level for these scenarios; and on the other hand experimentally, comparing passive and active RFID alternatives regarding costs, power consumption, distance ranges, collision problems, and ease of reconfiguration. Finally, the performance of the proposed on-board system is experimentally validated, testing its capabilities for inventory purposes.