Deploying deep learning approaches to left ventricular non-compaction measurement

The Journal of Supercomputing - Left ventricular non-compaction (LVNC) is a rare cardiomyopathy characterized by abnormal trabeculations in the left ventricle cavity. Although traditional computer...     

Integrating supercomputing clusters into education: a case study in biotechnology

The Journal of Supercomputing - The integration of a Supercomputer in the educational process improves student’s technological skills. The aim of the paper is to study the interaction between...     

How far can one compress digital fault records? Analysis of a matching pursuit-based algorithm

In this paper, we verify how far electric disturbance signals can be compressed without compromising the analysis of encoded fault records. A recently proposed compression algorithm, referred to as Damped Sinusoidal Matching Pursuit (DSMP) has the remarkable feature of obtaining both compact and physically interpretable representations. However, for fault analysis applications, one is primarily interested in how accurate can be the analysis performed on compressed signals, instead of evaluating mean-squared error figures. Unlike previous works in digital fault records compression, the performance of the DSMP compression method is evaluated using a protocol based on fault analysis procedures commonly performed by expert engineers. This protocol is applied for comparing the results obtained in the analysis of both uncompressed records and their compressed versions at different compression ratios. The results show that the DSMP is a reliable compression system since it achieves high compression ratios (6.4:1) without causing fault analysis misinterpretation.     

A novel low complexity data demodulation algorithm for pulse position modulation

In this paper, a low complexity data demodulation algorithm is proposed that requires time of arrival information of the received signal exclusively. As an application example, the algorithm is applied to an ultra-wideband impulse radio communication system with pulse position modulation. The algorithm is insensitive to a common time delay for all pulses, that means, it does not require an accurate synchronization between the transmitter and the receiver. For the performance estimation, only a symbol synchronization is assumed, i.e., that there is a priori knowledge which pulse marks the beginning of a received data symbol. The performance of the proposed algorithm is evaluated for straightforward time of arrivals estimators, such as a maximum detector or a threshold detector. It is shown that the proposed algorithm outperforms a least squares algorithm in all considered scenarios. In particular, an increased robustness against additive white Gaussian noise, impulse like noise, and multiuser interference is demonstrated as well as an improved performance for multipath propagation channels.     

A neuron membrane mesh representation for visualization of electrophysiological simulations

We present a process to automatically generate three-dimensional mesh representations of the complex, arborized cell membrane surface of cortical neurons (the principal information processing cells of the brain) from nonuniform morphological measurements. Starting from manually sampled morphological points (3D points and diameters) from neurons in a brain slice preparation, we construct a polygonal mesh representation that realistically represents the continuous membrane surface, closely matching the original experimental data. A mapping between the original morphological points and the newly generated mesh enables simulations of electrophysiolgical activity to be visualized on this new membrane representation. We compare the new mesh representation with the state of the art and present a series of use cases and applications of this technique to visualize simulations of single neurons and networks of multiple neurons.     

Action recognition via bio-inspired features: The richness of center–surround interaction

Motion is a key feature for a wide class of computer vision approaches to recognize actions. In this article, we show how to define bio-inspired features for action recognition. To do so, we start from a well-established bio-inspired motion model of cortical areas V1 and MT. The primary visual cortex, designated as V1, is the first cortical area encountered in the visual stream processing and early responses of V1 cells consist in tiled sets of selective spatiotemporal filters. The second cortical area of interest in this article is area MT where MT cells pool incoming information from V1 according to the shape and characteristic of their receptive field. To go beyond the classical models and following the observations from Xiao et al. [61], we propose here to model different surround geometries for MT cells receptive fields. Then, we define the so-called bio-inspired features associated to an input video, based on the average activity of MT cells. Finally, we show how these features can be used in a standard classification method to perform action recognition. Results are given for the Weizmann and KTH databases. Interestingly, we show that the diversity of motion representation at the MT level (different surround geometries), is a major advantage for action recognition. On the Weizmann database, the inclusion of different MT surround geometries improved the recognition rate from 63.01 ± 2.07% up to 99.26 ± 1.66% in the best case. Similarly, on the KTH database, the recognition rate was significantly improved with the inclusion of MT different surround geometries (from 47.82 ± 2.71% up to 92.44 ± 0.01% in the best case). We also discussed the limitations of the current approach which are closely related to the input video duration. These promising results encourage us to further develop bio-inspired models incorporating other brain mechanisms and cortical areas in order to deal with more complex videos.     

Evaluating the influence of multiplexing schemes and buffer implementation on perceived VoIP conversation quality

This work presents a study of RTP multiplexing schemes, which are compared with the normal use of RTP, in terms of experienced quality. Bandwidth saving, latency and packet loss for different options are studied, and some tests of Voice over IP (VoIP) traffic are carried out in order to compare the quality obtained using different implementations of the router buffer. Voice quality is calculated using ITU R-factor, which is a widely accepted quality estimator. The tests show the bandwidth savings of multiplexing, and also the importance of packet size for certain buffers, as latency and packet loss may be affected. The customer’s experience improvement is measured, showing that the use of multiplexing can be interesting in some scenarios, like an enterprise with different offices connected via the Internet. The system is also tested using different numbers of samples per packet, and the distribution of the flows into different tunnels is found to be an important factor in order to achieve an optimal perceived quality for each kind of buffer. Grouping all the flows into a single tunnel will not always be the best solution, as the increase of the number of flows does not improve bandwidth efficiency indefinitely. If the buffer penalizes big packets, it will be better to group the flows into a number of tunnels. The router processing capacity has to be taken into account too, as the limit of packets per second it can manage must not be exceeded. The obtained results show that multiplexing is a good way to improve customer’s experience of VoIP in scenarios where many RTP flows share the same path.     

Confidence voting method ensemble applied to off-line signature verification

In this paper, a new approximation to off-line signature verification is proposed based on two-class classifiers using an expert decisions ensemble. Different methods to extract sets of local and a global features from the target sample are detailed. Also a normalization by confidence voting method is used in order to decrease the final equal error rate (EER). Each set of features is processed by a single expert, and on the other approach proposed, the decisions of the individual classifiers are combined using weighted votes. Experimental results are given using a subcorpus of the large MCYT signature database for random and skilled forgeries. The results show that the weighted combination outperforms the individual classifiers significantly. The best EER obtained were 6.3 % in the case of skilled forgeries and 2.31 % in the case of random forgeries.     

Missing data imputation for fuzzy rule-based classification systems

Fuzzy rule-based classification systems (FRBCSs) are known due to their ability to treat with low quality data and obtain good results in these scenarios. However, their application in problems with missing data are uncommon while in real-life data, information is frequently incomplete in data mining, caused by the presence of missing values in attributes. Several schemes have been studied to overcome the drawbacks produced by missing values in data mining tasks; one of the most well known is based on preprocessing, formerly known as imputation. In this work, we focus on FRBCSs considering 14 different approaches to missing attribute values treatment that are presented and analyzed. The analysis involves three different methods, in which we distinguish between Mamdani and TSK models. From the obtained results, the convenience of using imputation methods for FRBCSs with missing values is stated. The analysis suggests that each type behaves differently while the use of determined missing values imputation methods could improve the accuracy obtained for these methods. Thus, the use of particular imputation methods conditioned to the type of FRBCSs is required.     

Towards practical anisotropic adaptive FEM on triangular meshes: a new refinement paradigm

Adaptive anisotropic refinement of finite element meshes allows one to reduce the computational effort required to achieve a specified accuracy of the solution of a PDE problem. We present a new approach to adaptive refinement and demonstrate that this allows one to construct algorithms which generate very flexible and efficient anisotropically refined meshes, even improving the convergence order compared to adaptive isotropic refinement if the problem permits.