Neurodegenerative diseases pose a formidable challenge to medical research, demanding a nuanced understanding of their progressive nature. In this regard, latent generative models can effectively be used in a data-driven modeling of different dimensions of neurodegeneration, framed within the context of the manifold hypothesis. This paper proposes a joint framework for a multi-modal, common latent generative model to address the need for a more comprehensive understanding of the neurodegenerative landscape in the context of Parkinson’s disease (PD). The proposed architecture uses coupled variational autoencoders (VAEs) to joint model a common latent space to both neuroimaging and clinical data from the Parkinson’s Progression Markers Initiative (PPMI). Alternative loss functions, different normalization procedures, and the interpretability and explainability of latent generative models are addressed, leading to a model that was able to predict clinical symptomatology in the test set, as measured by the unified Parkinson’s disease rating scale (UPDRS), with R2 up to 0.86 for same-modality and 0.441 cross-modality (using solely neuroimaging). The findings provide a foundation for further advancements in the field of clinical research and practice, with potential applications in decision-making processes for PD. The study also highlights the limitations and capabilities of the proposed model, emphasizing its direct interpretability and potential impact on understanding and interpreting neuroimaging patterns associated with PD symptomatology.

The combination of different sources of information is currently one of the most relevant aspects in the diagnostic process of several diseases. In the field of neurological disorders, different imaging modalities providing structural and functional information are frequently available. Those modalities are usually analyzed separately, although a joint of the features extracted from both sources can improve the classification performance of Computer-Aided Diagnosis (CAD) tools. Previous studies have computed independent models from each individual modality and combined them in a subsequent stage, which is not an optimum solution. In this work, we propose a method based on the principles of siamese neural networks to fuse information from Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET). This framework quantifies the similarities between both modalities and relates them with the diagnostic label during the training process. The resulting latent space at the output of this network is then entered into an attention module in order to evaluate the relevance of each brain region at different stages of the development of Alzheimer's disease. The excellent results obtained and the high flexibility of the method proposed allow fusing more than two modalities, leading to a scalable methodology that can be used in a wide range of contexts.

Parkinson's Disease (PD) is the second most prevalent neurodegenerative disorder among adults. Although its triggers are still not clear, they may be due to a combination of different types of biomarkers measured through medical imaging, metabolomics, proteomics or genetics, among others. In this context, we have proposed a Computer-Aided Diagnosis (CAD) system that combines structural and functional imaging data from subjects in Parkinson's Progression Markers Initiative dataset by means of an Ensemble Learning methodology trained to identify and penalize input sources with low classification rates and/ or high-variability. This proposal improves results published in recent years and provides an accurate solution not only from the point of view of image preprocessing (including a comparison between different intensity preservation techniques), but also in terms of dimensionality reduction methods (Isomap). In addition, we have also introduced a bagging classification schema for scenarios with unbalanced data. As shown by our results, the CAD proposal is able to detect PD with [Formula: see text] of balanced accuracy, and opens up the possibility of combining any number of input data sources relevant for PD.

The prevalence of dementia is currently increasing worldwide. This syndrome produces a deterioration in cognitive function that cannot be reverted. However, an early diagnosis can be crucial for slowing its progress. The Clock Drawing Test (CDT) is a widely used paper-and-pencil test for cognitive assessment in which an individual has to manually draw a clock on a paper. There are a lot of scoring systems for this test and most of them depend on the subjective assessment of the expert. This study proposes a computer-aided diagnosis (CAD) system based on artificial intelligence (AI) methods to analyze the CDT and obtain an automatic diagnosis of cognitive impairment (CI). This system employs a preprocessing pipeline in which the clock is detected, centered and binarized to decrease the computational burden. Then, the resulting image is fed into a Convolutional Neural Network (CNN) to identify the informative patterns within the CDT drawings that are relevant for the assessment of the patient's cognitive status. Performance is evaluated in a real context where patients with CI and controls have been classified by clinical experts in a balanced sample size of [Formula: see text] drawings. The proposed method provides an accuracy of [Formula: see text] in the binary case-control classification task, with an AUC of [Formula: see text]. These results are indeed relevant considering the use of the classic version of the CDT. The large size of the sample suggests that the method proposed has a high reliability to be used in clinical contexts and demonstrates the suitability of CAD systems in the CDT assessment process. Explainable artificial intelligence (XAI) methods are applied to identify the most relevant regions during classification. Finding these patterns is extremely helpful to understand the brain damage caused by CI. A validation method using resubstitution with upper bound correction in a machine learning approach is also discussed.

The automation in the diagnosis of medical images is currently a challenging task. The use of Computer Aided Diagnosis (CAD) systems can be a powerful tool for clinicians, especially in situations when hospitals are overflowed. These tools are usually based on artificial intelligence (AI), a field that has been recently revolutionized by deep learning approaches. These alternatives usually obtain a large performance based on complex solutions, leading to a high computational cost and the need of having large databases. In this work, we propose a classification framework based on sparse coding. Images are first partitioned into different tiles, and a dictionary is built after applying PCA to these tiles. The original signals are then transformed as a linear combination of the elements of the dictionary. Then, they are reconstructed by iteratively deactivating the elements associated with each component. Classification is finally performed employing as features the subsequent reconstruction errors. Performance is evaluated in a real context where distinguishing between four different pathologies: control versus bacterial pneumonia versus viral pneumonia versus COVID-19. Our system differentiates between pneumonia patients and controls with an accuracy of 97.74%, whereas in the 4-class context the accuracy is 86.73%. The excellent results and the pioneering use of sparse coding in this scenario evidence that our proposal can assist clinicians when their workload is high.

Spatial normalization helps us to compare quantitatively two or more input brain scans. Although using an affine normalization approach preserves the anatomical structures, the neuroimaging field is more common to find works that make use of nonlinear transformations. The main reason is that they facilitate a voxel-wise comparison, not only when studying functional images but also when comparing MRI scans given that they fit better to a reference template. However, the amount of bias introduced by the nonlinear transformations can potentially alter the final outcome of a diagnosis especially when studying functional scans for neurological disorders like Parkinson's Disease. In this context, we have tried to quantify the bias introduced by the affine and the nonlinear spatial registration of FP-CIT SPECT volumes of healthy control subjects and patients with PD. For that purpose, we calculated the deformation fields of each participant and applied these deformation fields to a 3D-grid. As the space between the edges of small cubes comprising the grid change, we can quantify which parts from the brain have been enlarged, compressed or just remain the same. When the nonlinear approach is applied, scans from PD patients show a region near their striatum very similar in shape to that of healthy subjects. This artificially increases the interclass separation between patients with PD and healthy subjects as the local intensity is decreased in the latter region, and leads machine learning systems to biased results due to the artificial information introduced by these deformations.

There remains an open question about the usefulness and the interpretation of machine learning (ML) approaches for discrimination of spatial patterns of brain images between samples or activation states. In the last few decades, these approaches have limited their operation to feature extraction and linear classification tasks for between-group inference. In this context, statistical inference is assessed by randomly permuting image labels or by the use of random effect models that consider between-subject variability. These multivariate ML-based statistical pipelines, whilst potentially more effective for detecting activations than hypotheses-driven methods, have lost their mathematical elegance, ease of interpretation, and spatial localization of the ubiquitous General linear Model (GLM). Recently, the estimation of the conventional GLM parameters has been demonstrated to be connected to an univariate classification task when the design matrix in the GLM is expressed as a binary indicator matrix. In this paper we explore the complete connection between the univariate GLM and ML-based regressions. To this purpose we derive a refined statistical test with the GLM based on the parameters obtained by a linear Support Vector Regression (SVR) in the inverse problem (SVR-iGLM). Subsequently, random field theory (RFT) is employed for assessing statistical significance following a conventional GLM benchmark. Experimental results demonstrate how parameter estimations derived from each model (mainly GLM and SVR) result in different experimental design estimates that are significantly related to the predefined functional task. Moreover, using real data from a multisite initiative the proposed ML-based inference demonstrates statistical power and the control of false positives, outperforming the regular GLM.

A connection between the general linear model (GLM) with frequentist statistical testing and machine learning (MLE) inference is derived and illustrated. Initially, the estimation of GLM parameters is expressed as a Linear Regression Model (LRM) of an indicator matrix; that is, in terms of the inverse problem of regressing the observations. Both approaches, i.e. GLM and LRM, apply to different domains, the observation and the label domains, and are linked by a normalization value in the least-squares solution. Subsequently, we derive a more refined predictive statistical test: the linear Support Vector Machine (SVM), that maximizes the class margin of separation within a permutation analysis. This MLE-based inference employs a residual score and associated upper bound to compute a better estimation of the actual (real) error. Experimental results demonstrate how parameter estimations derived from each model result in different classification performance in the equivalent inverse problem. Moreover, using real data, the MLE-based inference including model-free estimators demonstrates an efficient trade-off between type I errors and statistical power.

In recent years, several computer-aided diagnosis (CAD) systems have been proposed for an early identification of dementia. Although these approaches have mostly used the transformation of data into a different feature space, more precise information can be gained from a Searchlight strategy. The current study presents a data fusion classification system that employs magnetic resonance imaging (MRI) and neuropsychological tests to distinguish between Mild-Cognitive Impairment (MCI) patients that convert to Alzheimer’s disease (AD) and those that remain stable. Specifically, this method uses a nested cross-validation procedure to compute the optimum contribution of each data modality in the final decision. The model employs Support-Vector Machine (SVM) classifiers for both data modalities and is combined with Searchlight when applied to neuroimaging. We compared the performance of our system with an alternative based on Principal Component Analysis (PCA) for dimensionality reduction. Results show that Searchlight outperformed PCA both for uni/multimodal classification, obtaining a maximum accuracy of 80.9% when combining data from six and twelve months before patients converted to AD. Moreover, Searchlight allowed the identification of the most informative regions at different stages of the longitudinal study, which can be crucial for a better understanding of the development of AD. Additionally, results do not depend on the parcellations provided by a specific brain atlas, which manifests the robustness and the spatial precision of the method proposed.

Finding new biomarkers to model Parkinson's Disease (PD) is a challenge not only to help discerning between Healthy Control (HC) subjects and patients with potential PD but also as a way to measure quantitatively the loss of dopaminergic neurons mainly concentrated at substantia nigra. Within this context, this work presented here tries to provide a set of imaging features based on morphological characteristics extracted from I[Formula: see text]-Ioflupane SPECT scans to discern between HC and PD participants in a balanced set of [Formula: see text] scans from Parkinson's Progression Markers Initiative (PPMI) database. These features, obtained from isosurfaces of each scan at different intensity levels, have been classified through the use of classical Machine Learning classifiers such as Support-Vector-Machines (SVM) or Naïve Bayesian and compared with the results obtained using a Multi-Layer Perceptron (MLP). The proposed system, based on a Mann-Whitney-Wilcoxon U-Test for feature selection and the SVM approach, yielded a [Formula: see text] balanced accuracy when the performance was evaluated using a [Formula: see text]-fold cross-validation. This proves the reliability of these biomarkers, especially those related to sphericity, center of mass, number of vertices, 2D-projected perimeter or the 2D-projected eccentricity, among others, but including both internal and external isosurfaces.

The Temporal Sampling Framework (TSF) theorizes that the characteristic phonological difficulties of dyslexia are caused by an atypical oscillatory sampling at one or more temporal rates. The LEEDUCA study conducted a series of Electroencephalography (EEG) experiments on children listening to amplitude modulated (AM) noise with slow-rythmic prosodic (0.5-1[Formula: see text]Hz), syllabic (4-8[Formula: see text]Hz) or the phoneme (12-40[Formula: see text]Hz) rates, aimed at detecting differences in perception of oscillatory sampling that could be associated with dyslexia. The purpose of this work is to check whether these differences exist and how they are related to children's performance in different language and cognitive tasks commonly used to detect dyslexia. To this purpose, temporal and spectral inter-channel EEG connectivity was estimated, and a denoising autoencoder (DAE) was trained to learn a low-dimensional representation of the connectivity matrices. This representation was studied via correlation and classification analysis, which revealed ability in detecting dyslexic subjects with an accuracy higher than 0.8, and balanced accuracy around 0.7. Some features of the DAE representation were significantly correlated ([Formula: see text]) with children's performance in language and cognitive tasks of the phonological hypothesis category such as phonological awareness and rapid symbolic naming, as well as reading efficiency and reading comprehension. Finally, a deeper analysis of the adjacency matrix revealed a reduced bilateral connection between electrodes of the temporal lobe (roughly the primary auditory cortex) in DD subjects, as well as an increased connectivity of the F7 electrode, placed roughly on Broca's area. These results pave the way for a complementary assessment of dyslexia using more objective methodologies such as EEG.

Artificial intelligence and all its supporting tools, e.g. machine and deep learning in computational intelligence-based systems, are rebuilding our society (economy, education, life-style, etc.) and promising a new era for the social welfare state. In this paper we summarize recent advances in data science and artificial intelligence within the interplay between natural and artificial computation. A review of recent works published in the latter field and the state the art are summarized in a comprehensive and self-contained way to provide a baseline framework for the international community in artificial intelligence. Moreover, this paper aims to provide a complete analysis and some relevant discussions of the current trends and insights within several theoretical and application fields covered in the essay, from theoretical models in artificial intelligence and machine learning to the most prospective applications in robotics, neuroscience, brain computer interfaces, medicine and society, in general.

Many classical machine learning techniques have been used to explore Alzheimer's disease (AD), evolving from image decomposition techniques such as principal component analysis toward higher complexity, non-linear decomposition algorithms. With the arrival of the deep learning paradigm, it has become possible to extract high-level abstract features directly from MRI images that internally describe the distribution of data in low-dimensional manifolds. In this work, we try a new exploratory data analysis of AD based on deep convolutional autoencoders. We aim at finding links between cognitive symptoms and the underlying neurodegeneration process by fusing the information of neuropsychological test outcomes, diagnoses, and other clinical data with the imaging features extracted solely via a data-driven decomposition of MRI. The distribution of the extracted features in different combinations is then analyzed and visualized using regression and classification analysis, and the influence of each coordinate of the autoencoder manifold over the brain is estimated. The imaging-derived markers could then predict clinical variables with correlations above 0.6 in the case of neuropsychological evaluation variables such as the MMSE or the ADAS11 scores, achieving a classification accuracy over 80% for the diagnosis of AD.

Although much research has been undertaken, the spatial patterns, developmental course, and sexual dimorphism of brain structure associated with autism remains enigmatic. One of the difficulties in investigating differences between the sexes in autism is the small sample sizes of available imaging datasets with mixed sex. Thus, the majority of the investigations have involved male samples, with females somewhat overlooked. This paper deploys machine learning on partial least squares feature extraction to reveal differences in regional brain structure between individuals with autism and typically developing participants. A four-class classification problem (sex and condition) is specified, with theoretical restrictions based on the evaluation of a novel upper bound in the resubstitution estimate. These conditions were imposed on the classifier complexity and feature space dimension to assure generalizable results from the training set to test samples. Accuracies above [Formula: see text] on gray and white matter tissues estimated from voxel-based morphometry (VBM) features are obtained in a sample of equal-sized high-functioning male and female adults with and without autism ([Formula: see text], [Formula: see text]/group). The proposed learning machine revealed how autism is modulated by biological sex using a low-dimensional feature space extracted from VBM. In addition, a spatial overlap analysis on reference maps partially corroborated predictions of the “extreme male brain” theory of autism, in sexual dimorphic areas.

Medical image classification is currently a challenging task that can be used to aid the diagnosis of different brain diseases. Thus, exploratory and discriminative analysis techniques aiming to obtain representative features from the images play a decisive role in the design of effective Computer Aided Diagnosis (CAD) systems, which is especially important in the early diagnosis of dementia. In this work, we present a technique that allows using specific time series analysis techniques with 3D images. This is achieved by sampling the image using a fractal-based method which preserves the spatial relationship among voxels. In addition, a method called Empirical functional PCA (EfPCA) is presented, which combines Empirical Mode Decomposition (EMD) with functional PCA to express an image in the space spanned by a basis of empirical functions, instead of using components computed by a predefined basis as in Fourier or Wavelet analysis. The devised technique has been used to classify images from the Alzheimer's Disease Neuroimaging Initiative (ADNI) and the Parkinson Progression Markers Initiative (PPMI), achieving accuracies up to 93% and 92% differential diagnosis tasks (AD versus controls and PD versus Controls, respectively). The results obtained validate the method, proving that the information retrieved by our methodology is significantly linked to the diseases.

Parkinsonism is a clinical syndrome characterized by the progressive loss of striatal dopamine. Its diagnosis is usually corroborated by neuroimaging data such as DaTSCAN neuroimages that allow visualizing the possible dopamine deficiency. During the last decade, a number of computer systems have been proposed to automatically analyze DaTSCAN neuroimages, eliminating the subjectivity inherent to the visual examination of the data. In this work, we propose a computer system based on machine learning to separate Parkinsonian patients and control subjects using the size and shape of the striatal region, modeled from DaTSCAN data. First, an algorithm based on adaptative thresholding is used to parcel the striatum. This region is then divided into two according to the brain hemisphere division and characterized with 152 measures, extracted from the volume and its three possible 2-dimensional projections. Afterwards, the Bhattacharyya distance is used to discard the least discriminative measures and, finally, the neuroimage category is estimated by means of a Support Vector Machine classifier. This method was evaluated using a dataset with 189 DaTSCAN neuroimages, obtaining an accuracy rate over 94%. This rate outperforms those obtained by previous approaches that use the intensity of each striatal voxel as a feature.

Spatial and intensity normalizations are nowadays a prerequisite for neuroimaging analysis. Influenced by voxel-wise and other univariate comparisons, where these corrections are key, they are commonly applied to any type of analysis and imaging modalities. Nuclear imaging modalities such as PET-FDG or FP-CIT SPECT, a common modality used in Parkinson's disease diagnosis, are especially dependent on intensity normalization. However, these steps are computationally expensive and furthermore, they may introduce deformations in the images, altering the information contained in them. Convolutional neural networks (CNNs), for their part, introduce position invariance to pattern recognition, and have been proven to classify objects regardless of their orientation, size, angle, etc. Therefore, a question arises: how well can CNNs account for spatial and intensity differences when analyzing nuclear brain imaging? Are spatial and intensity normalizations still needed? To answer this question, we have trained four different CNN models based on well-established architectures, using or not different spatial and intensity normalization preprocessings. The results show that a sufficiently complex model such as our three-dimensional version of the ALEXNET can effectively account for spatial differences, achieving a diagnosis accuracy of 94.1% with an area under the ROC curve of 0.984. The visualization of the differences via saliency maps shows that these models are correctly finding patterns that match those found in the literature, without the need of applying any complex spatial normalization procedure. However, the intensity normalization - and its type - is revealed as very influential in the results and accuracy of the trained model, and therefore must be well accounted.

Computer-aided diagnosis (CAD) systems constitute a powerful tool for early diagnosis of Alzheimer's disease (AD), but limitations on interpretability and performance exist. In this work, a fully automatic CAD system based on supervised learning methods is proposed to be applied on segmented brain magnetic resonance imaging (MRI) from Alzheimer's disease neuroimaging initiative (ADNI) participants for automatic classification. The proposed CAD system possesses two relevant characteristics: optimal performance and visual support for decision making. The CAD is built in two stages: a first feature extraction based on independent component analysis (ICA) on class mean images and, secondly, a support vector machine (SVM) training and classification. The obtained features for classification offer a full graphical representation of the images, giving an understandable logic in the CAD output, that can increase confidence in the CAD support. The proposed method yields classification results up to 89% of accuracy (with 92% of sensitivity and 86% of specificity) for normal controls (NC) and AD patients, 79% of accuracy (with 82% of sensitivity and 76% of specificity) for NC and mild cognitive impairment (MCI), and 85% of accuracy (with 85% of sensitivity and 86% of specificity) for MCI and AD patients.

This paper shows a magnetic resonance image (MRI) classification technique based on nonnegative matrix factorization (NNMF) and ensemble tree learning methods. The system consists of a feature extraction process that applies NNMF to gray matter (GM) MRI first-order statistics of a number of sub-cortical structures and a learning process of an ensemble of decision trees. The ensembles are trained by means of boosting and bagging while their performance is compared in terms of the classification error and the received operating characteristics curve (ROC) using k-fold cross validation. The results show that NNMF is well suited for reducing the dimensionality of the input data without a penalty on the performance of the ensembles. The best performance was obtained by bagging in terms of convergence rate and minimum residual loss, especially for high complexity classification tasks (i.e. NC vs. MCI and MCI vs. AD.

The usage of biomedical imaging in the diagnosis of dementia is increasingly widespread. A number of works explore the possibilities of computational techniques and algorithms in what is called computed aided diagnosis. Our work presents an automatic parametrization of the brain structure by means of a path generation algorithm based on hidden Markov models (HMMs). The path is traced using information of intensity and spatial orientation in each node, adapting to the structure of the brain. Each path is itself a useful way to characterize the distribution of the tissue inside the magnetic resonance imaging (MRI) image by, for example, extracting the intensity levels at each node or generating statistical information of the tissue distribution. Additionally, a further processing consisting of a modification of the grey level co-occurrence matrix (GLCM) can be used to characterize the textural changes that occur throughout the path, yielding more meaningful values that could be associated to Alzheimer's disease (AD), as well as providing a significant feature reduction. This methodology achieves moderate performance, up to 80.3% of accuracy using a single path in differential diagnosis involving Alzheimer-affected subjects versus controls belonging to the Alzheimer's disease neuroimaging initiative (ADNI).

Computer Aided Diagnosis (CAD) constitutes an important tool for the early diagnosis of Alzheimer's Disease (AD), which, in turn, allows the application of treatments that can be simpler and more likely to be effective. This paper explores the construction of classification methods based on deep learning architectures applied on brain regions defined by the Automated Anatomical Labeling (AAL). Gray Matter (GM) images from each brain area have been split into 3D patches according to the regions defined by the AAL atlas and these patches are used to train different deep belief networks. An ensemble of deep belief networks is then composed where the final prediction is determined by a voting scheme. Two deep learning based structures and four different voting schemes are implemented and compared, giving as a result a potent classification architecture where discriminative features are computed in an unsupervised fashion. The resulting method has been evaluated using a large dataset from the Alzheimer's disease Neuroimaging Initiative (ADNI). Classification results assessed by cross-validation prove that the proposed method is not only valid for differentiate between controls (NC) and AD images, but it also provides good performances when tested for the more challenging case of classifying Mild Cognitive Impairment (MCI) Subjects. In particular, the classification architecture provides accuracy values up to 0.90 and AUC of 0.95 for NC/AD classification, 0.84 and AUC of 0.91 for stable MCI/AD classification and 0.83 and AUC of 0.95 for NC/MCI converters classification.

This work highlights the exploitation of Gaussian Mixture Model (GMM) and Mean squared Error (MSE) in DaTSCAN SPECT brain images for intensity normalization purposes over two proposed approaches. The first proposed methodology is based on a nonlinear image filtering by means of GMM, which considers not only the intensity levels of each voxel but also its coordinates inside the so-defined spatial Gaussian functions. It is achieved according to a probability threshold that measures the weight of each kernel or cluster on the striatum area, the voxels in the non-specific regions are intensity normalized by removing clusters whose likelihood is negligible. The second normalization method based on MSE which is performed by a linear intensity transformation in each voxel. This approach is based on predicting jointly different intensity normalization parameters that leads to the joint minimization of the squared sum errors between the template image and the optimal linear estimated image (normalized image). We compare these methods of normalization together with another approach widely used based on specific-to-non-specific binding ratio. This comparison is based on DaTSCAN image analysis and classification for the development of a computer aided diagnosis (CAD) system for Parkinsonian syndrome detection.

This work presents a novel feature selection method for Positron Emission Tomography (PET) images for the development of a computer aided diagnosis (CAD) system aiming to improve the early detection of the Alzheimer’s Disease (AD). Voxels are selected by means of the combination of the antecedents and consequents of Association Rules (ARs) which are mined from the Regions of Interest (ROIs) of controls by Apriori algorithm. In order to reduce the input space to the classifier, features are extracted using Principal Component Analysis (PCA) or Partial Least Squares (PLS) techniques. Finally, classification is performed by using a kernel Support Vector Machine (SVM) reaching accuracy rates of 90% outperforming other reported methods.

In this paper we propose a novel method for brain SPECT image feature extraction based on the Empirical Mode Decomposition (EMD). The proposed method applied to assist the diagnosis of Alzheimer Disease (AD) selects the most discriminant voxels for classification from the transformed EMD feature space. In particular, high-frequency components of the EMD transformation are found to retain regional differences in functional activity which is characteristic of AD. The EMD represents a fully data-driven, unsupervised and additive signal decomposition and does not need any a priori defined basis system. Several experiments were carried out on a balanced SPECT database collected from the “Virgen de las Nieves” Hospital in Granada (Spain), containing 96 recordings and yielding up to 100% accuracy in separating AD and normal controls, and then outperforming recently proposed CAD systems.

Image segmentation aims to partition an image into its constituent parts. In Magnetic Resonance Imaging (MRI), image segmentation groups the voxels belonging to a specific tissue or fluids. Currently, this is an essential step in MRI analysis and provides a way to diagnose many neurological disorders such as the Alzheimer disease. In this paper we present a segmentation method for MRI based on the Growing Hierarchical Self-Organizing Map and multiobjective-based feature selection to optimize the performance of the segmentation process. Since the features extracted from the image result crucial for the final performance of the segmentation process, the optimized features which maximize the performance of the segmentation process on each plane are used. The experiments performed have been carried out using the Internet Brain Segmentation Repository (IBSR) as it has been widely used in previous works. Thus, the comparisons with other methods carried out using the IBSR database, show that our method outperforms other algorithms.

Imaging of dopamine transporters (DaT) has been introduced as a valuable tool to evaluate patients with several neuropsychiatrie disorders, such as Parkinson’s disease (PD). Over the last decade, several computer applications have been developed in order to facilitate the exploration of DaT images to clinicians, however they require a high interaction with the user to delimit regions of interest (ROIs) and to obtain a diagnosis. In this paper we show a full automatic computed aided diagnosis (CAD) system for Parkinson’s disease. We use a Partial Least Square scheme to decompose DaT images into scores and loading. Then, the scores with highest Fisher Discriminant Ratio are used as feature for a Support Vector Machine classifier that determines the state (normal of pathological) of a given DaT image. The evaluation of the system using a database with 178 images from controls and PD patients has produced accuracy rates of nearly 93%.

This paper presents a novel computer-aided diagnosis (CAD) technique for the early diagnosis of the Alzheimer's disease (AD) based on nonnegative matrix factorization (NMF) and support vector machines (SVM) with bounds of confidence. The CAD tool is designed for the study and classification of functional brain images. For this purpose, two different brain image databases are selected: a single photon emission computed tomography (SPECT) database and positron emission tomography (PET) images, both of them containing data for both Alzheimer's disease (AD) patients and healthy controls as a reference. These databases are analyzed by applying the Fisher discriminant ratio (FDR) and nonnegative matrix factorization (NMF) for feature selection and extraction of the most relevant features. The resulting NMF-transformed sets of data, which contain a reduced number of features, are classified by means of a SVM-based classifier with bounds of confidence for decision. The proposed NMF-SVM method yields up to 91% classification accuracy with high sensitivity and specificity rates (upper than 90%). This NMF-SVM CAD tool becomes an accurate method for SPECT and PET AD image classification.

This paper presents a computer-aided diagnosis technique for improving the accuracy of the early diagnosis of the Alzheimer type dementia. The proposed methodology is based on the calculation of the skewness to each m-by-m sliding block of the transaxial slices of the SPECT brain images. We replace the center pixel in the m-by-m block by the skewness value and build a new 3-D brain image which will be used for classification purposes. After that, we select the voxels which present a Welch's t-statistic between both classes, Normal and Alzheimer images, higher (or lower) than a given threshold. The mean, standard deviation, skewness and kurtosis are calculated for selected voxels and they are chosen as feature vectors for three different classifiers: support vector machines with linear kernel, classification trees and multivariate normal model. The proposed methodology reaches an accuracy higher than 98% in the classification task.

We present a method for automatic seismic event detection and classification, focusing on volcanic-seismic signals by means of the validity of the hidden Markov modeling (HMM) method in active volcanoes. Recordings of different seismic event types are studied at one active volcano; San Cristobal in Nicaragua. We use data from one field surveys carried out in February to March 2006. More than 600 hours of data in San Cristobal volcano were analyzed and 1098 seismic events were registered at short period stations. These events were manually labelled by a single expert technicians and identified three types classes of signals (S1, S2, S3) and tremor background seismic noise (NS). The method analyzes the seismograms comparing the characteristics of the data to a number of event classes defined beforehand. If a signal is present, the method detects its occurrence and produces a classification. From the application performed over our data set, we have demonstrated that in order to have a reliable result, a careful and adequate segmentation process is crucial. Also, each type of signals requires its own characterization. That is, each signal type must be represented by its own specific model, which would include the effects of source, path and sites. Once we have built this model, the success level of the system is high. Extensive performance evaluation is conducted to derive the optimal configuration of the different parameters Correct classification rates of up to 80% are achieved. The high success rates obtained imply that the method is fully able to detect, isolate, and identify seismic signals on raw seismic data. These results imply that, once an adequate training process has been used, the present method is particularly appropriate to work in real time, and in parallel to the data acquisition.

In this letter, we propose a novel least-mean-square (LMS) algorithm for filtering speech sounds in the adaptive noise cancellation (ANC) problem. It is based on the minimization of the squared Euclidean norm of the difference weight vector under a stability constraint defined over theaposterioriestimation error. To this purpose, the Lagrangian methodology has been used in order to propose a nonlinear adaptation rule defined in terms of the product of differential inputs and errors which means a generalization of the normalized (N)LMS algorithm. The proposed method yields better tracking ability in this context as shown in the experiments which are carried out on the AURORA 2 and 3 speech databases. They provide an extensive performance evaluation along with an exhaustive comparison to standard LMS algorithms with almost the same computational load, including the NLMS and other recently reported LMS algorithms such as the modified (M)-NLMS, the error nonlinearity (EN)-LMS, or the normalized data nonlinearity (NDN)-LMS adaptation.

This paper presents a novel voice activity detector (VAD) for improving speech detection robustness in noisy environments and the performance of speech recognition systems in real-time applications. The algorithm is based on a generalized complex Gaussian (GCG) observation model and defines an optimal likelihood ratio test (LRT) involving multiple and correlated observations (MCO) based on jointly Gaussian probability distribution functions (jGpdf). An extensive analysis of the proposed methodology for a low dimensional observation model demonstrates 1) the improved robustness of the proposed approach by means of a clear reduction of the classification error as the number of observations is increased, and 2) the tradeoff between the number of observations and the detection performance. The proposed strategy is also compared to different VAD methods including the G.729, AMR, and AFE standards, as well as other recently reported algorithms showing a sustained advantage in speech/nonspeech detection accuracy and speech recognition performance.

This paper shows a revised statistical test for voice activity detection in noise adverse environments. The method is based on a revised contextual likelihood ratio test (LRT) defined over a multiple observation window. The new approach not only evaluates the two hypothesis consisting on all the observations to be speech or nonspeech but all the possible hypothesis defined over the individual observations. The implicit hangover mechanism artificially added by the original method was not found in the revised method so its design can be further improved. With these and other innovations the proposed method showed a high speech/non-speech discrimination over a wide range of SNR conditions. The experimental framework showed that the revised method yields significant improvements over standardized VADs for discontinuous voice transmission and distributed speech recognition, as well as over recently reported methods.

This paper shows an improved statistical test for voice activity detection in noise adverse environments. The method is based on a revised contextual likelihood ratio test (LRT) defined over a multiple observation window. The motivations for revising the original multiple observation LRT (MO-LRT) are found in its artificially added hangover mechanism that exhibits an incorrect behavior under different signal-to-noise ratio (SNR) conditions. The new approach defines a maximum a posteriori (MAP) statistical test in which all the global hypotheses on the multiple observation window containing up to one speech-to-nonspeech or nonspeech-to-speech transitions are considered. Thus, the implicit hangover mechanism artificially added by the original method was not found in the revised method so its design can be further improved. With these and other innovations, the proposed method showed a higher speech/nonspeech discrimination accuracy over a wide range of SNR conditions when compared to the original MO-LRT voice activity detector (VAD). Experiments conducted on the AURORA databases and tasks showed that the revised method yields significant improvements in speech recognition performance over standardized VADs such as ITU T G.729 and ETSI AMR for discontinuous voice transmission and the ETSI AFE for distributed speech recognition (DSR), as well as over recently reported methods.

This paper shows a complete seismic-event classification and monitoring system that has been developed based on the seismicity observed during three summer Antarctic surveys at the Deception Island Volcano, Antarctica. The system is based on the state of the art in hidden Markov modeling (HMM) techniques successfully applied to other scenarios. A database that contains a representative set of different seismic events including volcano-tectonic earthquakes, long period (LP) events, volcanic tremor, and hybrid events that were recorded during the 1994–1995 and 1995–1996 seismic surveys was collected for training and testing. Simple left-to-right HMMs and multivariate Gaussian probability density functions with a diagonal covariance matrix were used. The feature vector consists of the log-energies of a filter bank that consists of 16 triangular weighting functions that were uniformly spaced between 0 and 20 Hz and the first- and second-order derivatives. The system is suitable to operate in real time, and its accuracy for this task is about 90%. On the other hand, when the system was tested with a different data set including mainly LP events that were registered during several seismic swarms during the 2001–2002 field survey, more than 95% of the recognized events were marked by the recognition system.

This paper shows a complete volcano monitoring system that has been developed on the basis of the seismicity observed during three summer Antarctic surveys at Deception Island Volcano (Antarctica). The system is based on the state of the art in hidden Markov modelling (HMM) techniques successfully applied to other scenarios. A database containing a representative set of different seismic events including volcano-tectonic earthquakes, long-period events, volcanic tremor and hybrid events recorded during the 1994-1995 and 1995-1996 seismic surveys was collected for training and testing. Simple left-to-right HMMs and multivariate Gaussian probability density functions (PDF) with diagonal covariance matrix were used. The feature vector consists of the log-energies of a filter-bank consisting of 16 triangular weighting functions uniformly spaced between 0 and 20 Hz plus the first and second order derivatives. The system is suitable to operate in real-time and its accuracy is close to 90%. When the system was tested with a different data set including mainly long-period events registered during several seismic swarms during the 2001-2002 field survey, more than 95% of the recognized events were correctly marked by the recognition system

This paper shows an effective speech/pause discrimination method combining spectral noise filtering and fuzzy logic rules. The fuzzy system is based on a Sugeno inference engine with membership functions defined as combination of two Gaussian functions. Its operation is optimized by means of a hybrid training algorithm combining the least-squares method and the backpropagation gradient descent method for training membership function parameters. The fuzzy classifier consists of ten fuzzy rules defined in terms of the denoised subband signal-to-noise ratios (SNRs) and the zero crossing rate (ZCRs). An exhaustive analysis conducted on the Spanish SpeechDat-Car databases is conducted in order to assess the performance of the proposed method and to compare it to existing standard VAD methods. The results show improvements in detection accuracy over standard VADs and a representative set of recently reported VAD algorithms

A new front-end normalization algorithm that uses a parametric nonlinear transformation is proposed in this paper. The method improves histogram equalization based nonlinear transformations by finding a simple and computationally inexpensive parametric expression of the nonlinear transformation. The new parametric approach relies on a two Gaussian model for the probability distribution of the features, and on a simple Gaussian classifier to label the input frames as belonging to the speech or non-speech classes. The result is a more robust equalization, less dependent on the percentage of speech and non-speech frames. Recognition experiments on the AURORA 4 database have been performed and the effectiveness of the algorithm is analyzed in comparison with other linear and nonlinear feature equalization techniques

This paper shows an effective voice activity detector based on a statistical likelihood ratio test defined on the integrated bispectrum of the signal. It inherits the ability of higher order statistics to detect signals in noise with many other additional advantages: i) its computation as a cross spectrum leads to significant computational savings, and ii) the variance of the estimator is of the same order as that of the power spectrum estimator. The proposed method incorporates contextual information to the decision rule, a strategy that has reported significant improvements in speech detection accuracy and robust speech recognition applications. The experimental analysis conducted on the well-known AURORA databases has reported significant improvements over standardized techniques such as ITU G.729, AMRI, AMR2 and ESTI AFE VADs, as well as over recently published VADs

This paper presents GAIALAB, a Weblab project for digital communications distributed learning developed at the University of Granada (Spain). This initiative has been funded by a University program to improve teaching quality. The project addresses problems of teaching and learning innovatively and effectively by enabling the user interaction with the simulation engines through a Web browser. Our Weblab project provides an interface with JAVA applets that have been developed and integrated in the system. The Weblab project enables simulating digital communication systems and analyzing the results obtained through the Web. Moreover, all the applets that are included in the system have been designed to allow the modification of the simulation parameters, thus enabling a better understanding of the algorithms. The system also permits evaluating the experimental work developed by the students

A filter that introduces inter-frame information into the voice features set is proposed in this paper. The filter adds the autocorrelations of the cepstral coefficients to the set of characteristics used for training and recognition. Those autocorrelations should not depend on the environment conditions. Because they should only depend on the information to recognize, a normalization of that inter-frame information is convenient. The filter defined implements this normalization by transforming the autocorrelations into a normalized domain defined with clean adaptation data. This temporal processing of the features is added to the Histogram Equalization of the cepstral coefficients (HEQ) used to normalize the MFCCs. An analysis is done about the most effective domain (original MFCCS or equalized MFCCs) on which the temporal processing should be executed. Performance results for the proposed algorithm are presented for AURORA2 and AURORA4 databases.

Abstract In this paper we propose a novel Voice Activity Detection (VAD)algorithm, based on the integrated bispectrum function (IBI), forimproving Automated Speech Recognition (ASR) systems thatwork in noisy environments. In particular we use the combina-tion of two features, IBI magnitude and IBI phase to formulatea robust and smoothed decision rule for speech/pause discrimina-tion. The analysis performed on the new combined feature high-lighted: i) the advantages of each individual feature, while com-pensating the drawback of each other, and ii) the higher ability forendpoint detection given by a lower variance of the decision func-tion in pause/speech frames. The experiments conducted on theSpanish SpeechDat-Car database showed that the proposed algo-rithm outperforms ITU G.729, ETSI AMR1 and AMR2 and ETSIAFE standards as well as other recently reported VAD methods inspeech/non-speech detection performance.IndexTerms: voiceactivitydetection, clusteringanalysis, bispec-trum function, entropy.

This paper shows an effective speech/non-speech discrimination method for improving the performance of speech processing systems working in noisy environment. The proposed method uses a trained support vector machine (SVM) that defines an optimized non-linear decision rule over different sets of speech features. Two alternative feature extraction processes based on: i) subband SNR estimation after denoising, and ii) long-term SNR estimation were compared. Both methods show the ability of the SVM-based classifier to learn how the signal is masked by the acoustic noise and to define an effective non-linear decision rule. However, it is shown that a feature vector incorporating contextual information yielded better speech/non-speech discrimination even when no denoising is applied. The experimental analysis carried out on the Spanish SpeechDat-Car database shows clear improvements over standard VADs including ITU G.729, ETSI AMR and ETSI AFE for distributed speech recognition (DSR), and other recently reported VADs. Index Terms: voice activity detection, support vector machine learning, speech enhancement.

We propose a model-based VAD derived from the Vector Taylor Series (VTS) approach. A Gaussian mixture (trained with clean speech) is used in order to provide an appropriate decision rule for speech/non-speech detection. Additionally, VTS approach adapts the Gaussian mixture to noise conditions, yielding a stable perfor- mance for a wide range of SNRs. We have evaluated its ability for speech/non-speech detection and also its application for robust speech recognition. When compared to other VAD methods, the proposed VAD shows the best trade-off in speech/non-speech de- tection. When applied for Wiener Filtering and for frame drop- ping, the proposed VAD also provides the best recognition results. Index Terms: voice activity detection (VAD), vector Taylor series approach (VTS), Gaussian mixture, Wiener filtering.

A robust and effective voice activity detection (VAD) algorithm is proposed for improving speech recognition performance in noisy environments. The approach is based on well-known statistical tests based on the determination of the speech/non-speech bispectra by means of third-order auto-cumulants. This algorithm differs from many others in the way the decision rule is formulated being the statistical tests built on a multiple observation (MO) window consisting of averaged bispectrum coefficients of the speech signal. Clear improvements in speech/non-speech discrimination accuracy demonstrate the effectiveness of the proposed VAD. It is shown that application of a statistical detection test leads to a better separation of the speech and noise distributions, thus allowing a more effective discrimination and a tradeoff between complexity and performance. The experimental analysis carried out on the AURORA 3 databases provides an extensive performance evaluation together with an exhaustive comparison to the standard VADs, such as ITU G.729, GSM AMR, and ETSI AFE, for distributed speech recognition (DSR) and other recently reported VADs

This letter shows an effective statistical voice activity detection algorithm based on the integrated bispectrum, which is defined as a cross spectrum between the signal and its square and inherits the ability of higher order statistics to detect signals in noise with many other additional advantages: 1) its computation as a cross spectrum leads to significant computational savings, and 2) the variance of the estimator is of the same order as that of the power spectrum estimator. The decision rule is formulated in terms of an average likelihood ratio test (LRT) involving successive integrated bispectrum speech features. With these and other innovations, the proposed method reports significant improvements in speech/pause discrimination as well as in speech recognition over standardized techniques such as ITU-T G.729, ETSI AMR, and AFE VADs, and over recently published VADs

The discrete wavelet transform is routinely used in numerous applications and its efficient implementation in VLSI technology is a subject of continuous research. This paper shows the implementation of reduced complexity and high performance polyphase wavelet filter banks using Galois field arithmetic. The proposed designs are compared to traditional systems using 2's complement (2C) arithmetic with the proposed architectures being about 65% faster than 2C designs and requiring less logic in most study cases.

Currently, there are technology barriers inhibiting speech processing systems that work in extremely noisy conditions from meeting the demands of modern applications. This letter presents a new voice activity detector (VAD) for improving speech detection robustness in noisy environments and the performance of speech recognition systems. The algorithm defines an optimum likelihood ratio test (LRT) involving multiple and independent observations. The so-defined decision rule reports significant improvements in speech/nonspeech discrimination accuracy over existing VAD methods that are defined on a single observation and need empirically tuned hangover mechanisms. The algorithm has an inherent delay that, for several applications, including robust speech recognition, does not represent a serious implementation obstacle. An analysis of the overlap between the distributions of the decision variable shows the improved robustness of the proposed approach by means of a clear reduction of the classification error as the number of observations is increased. The proposed strategy is also compared to different VAD methods, including the G.729, AMR, and AFE standards, as well as recently reported algorithms showing a sustained advantage in speech/nonspeech detection accuracy and speech recognition performance.

An effective voice activity detection (VAD) algorithm is proposed for improving speech recognition performance in noisy environments. The approach is based on the determination of the speech/nonspeech divergence by means of specialized order statistics filters (OSFs) working on the subband log-energies. This algorithm differs from many others in the way the decision rule is formulated. Instead of making the decision based on the current frame, it uses OSFs on the subband log-energies which significantly reduces the error probability when discriminating speech from nonspeech in a noisy signal. Clear improvements in speech/nonspeech discrimination accuracy demonstrate the effectiveness of the proposed VAD. It is shown that an increase of the OSF order leads to a better separation of the speech and noise distributions, thus allowing a more effective discrimination and a tradeoff between complexity and performance. The algorithm also incorporates a noise reduction block working in tandem with the VAD and showed to further improve its accuracy. A previous noise reduction block also improves the accuracy in detecting speech and nonspeech. The experimental analysis carried out on the AURORA databases and tasks provides an extensive performance evaluation together with an exhaustive comparison to the standard VADs such as ITU G.729, GSM AMR, and ETSI AFE for distributed speech recognition (DSR), and other recently reported VADs.

Currently, there are technology barriers inhibiting speech processing systems working under extreme noisy conditions. The emerging applications of speech technology, especially in the fields of wireless communications, digital hearing aids or speech recognition, are some examples of such systems often requiring a noise reduction technique in combination with a precise voice activity detector (VAD). This paper presents a new VAD for improving speech detection robustness in noisy environments and the performance of speech recognition systems. The algorithm uses long-term information about the speech signal to formulate the decision rule and estimates the subband SNR using specialized order statistics filters (OSF). The proposed algorithm is compared to the most commonly used VAD in the field, in terms of speech/nonspeech discrimination and also in terms of recognition performance when the VAD is used in an automatic speech recognition (ASR) system. Experimental results demonstrate a sustained advantage over different VAD methods including standard VAD such as G.729 and AMR which are used as a reference, the VAD of the Advanced Front-End (AFE) for distributed speech recognition (DSR), and recently reported algorithms.

The paper mainly focusses on an improved voice activity detection algorithm employing long-term signal processing and maximum spectral component tracking. The benefits of this approach have been analyzed in a previous work (Ramirez, J. et al., Proc. EUROSPEECH 2003, p.3041-4, 2003) with clear improvements in speech/non-speech discriminability and speech recognition performance in noisy environments. Two clear aspects are now considered. The first one, which improves the performance of the VAD in low noise conditions, considers an adaptive length frame window to track the long-term spectral components. The second one reduces misclassification errors in highly noisy environments by using a noise reduction stage before the long-term spectral tracking. Experimental results show clear improvements over different VAD methods in speech/pause discrimination and speech recognition performance. Particularly, improvements in recognition rate were reported when the proposed VAD replaced the VADs of the ETSI advanced front-end (AFE) for distributed speech recognition (DSR).

Currently, new trends in wireless communications are demanding reliable human-machine interaction in real-life environments. However, there are obstacles inhibiting automatic speech recognition systems (ASR) working in noisy environments. The main difficulty is the degradation suffered by ASR systems due to a mismatch between training and test conditions. This paper shows an improved voice activity detector (VAD) combining noise reduction and subband divergence estimation for improving the reliability of speech recognizers operating in noisy environments. The algorithm formulates the decision rule by measuring the divergence between the subband spectral magnitude of speech and noise using the Kullback-Leibler (KL) distance on the denoised signal. Experiments demonstrate a sustained advantage over different VAD methods including standard VADs such as G.729 and AMR, which are used as a reference, recently reported algorithms, and the VADs of the advanced frontend (AFE) for distributed speech recognition (DSR). 1.

Noise compensation methods for speech recognition provide a cleaned version of the speech representation. Usually this cleaned version is the expected value of the speech parameters given the observed noisy speech and the noise statistic. A more realistic representation should include the probability distribution of the cleaned speech instead of its expected value in order to represent the uncertainty associated to the compensation process due to the variability of the noise process. Recently, the inclusion of the uncertainty in the recognition process has been studied. Some approaches represent the uncertainty in the HMM parameters values. Other approaches represent it in the feature space. This second approach offers a much simpler system implementation and lower computational cost. In this paper we have developed a noise compensation technique that incorporates the variance of the cleaned speech into the speech representation. The variance is estimated using a Wiener filter during the speech feature enhancement process. This way of including the uncertainty implies the modification of the decoding rule. Experimental results using AURORA 2 database demonstrate a sustained improvement of the performance in the recognition system (about 21% word error rate reduction) when uncertainty is considered in the decoding rule.

This letter shows an innovative voice activity detector (VAD) based on the Kullback-Leibler (KL) divergence measure. The algorithm is evaluated in the context of the recently approved ETSI standard for distributed speech recognition (DSR). The VAD uses long-term information of the noisy speech signal in order to define a more robust decision rule yielding high accuracy. The mel-scaled filter bank log-energies (FBE) are modeled by means of Gaussian distributions, and a symmetric KL divergence is used for the estimation of the distance between speech and noise distributions. The decision rule is formulated in terms of the average subband KL divergence that is compared to a noise-adaptable threshold. An exhaustive analysis using the AURORA databases is conducted in order to assess the performance of the proposed method and to compare it to existing standard VAD methods.

This letter presents a new segmental nonlinear feature normalization algorithm to improve the robustness of speech recognition systems against variations of the acoustic environment. An experimental study of the best delay-performance tradeoff is conducted within the AURORA-2 framework, and a comparison with two commonly used normalization algorithms is presented. Computationally efficient algorithms based on order statistics are also presented. One of them is based on linear interpolation between sampling quantiles, and the other one is based on a point estimation of the probability distribution. The reduction in the computational cost does not degrade the performance significantly.

This paper shows an efficient voice activity detector (VAD) that is based on the estimation of the long-term spectral diver- gence (LTSD) between noise and speech periods. The proposed method decomposes the input signal into overlapped speech frames, uses a sliding window to compute the long-term spec- tral envelope and measures the speech/non-speech LTSD, thus yielding a high discriminating decision rule and minimizing the average number of decision errors. In order to increase non- speech detection accuracy, the decision threshold is adapted to the measured noise energy while a controlled hang-over is ac- tivated only when the observed signal-to-noise ratio (SNR) is low. An exhaustive analysis of the proposed VAD is carried out using the AURORA TIdigits and SpeechDat-Car (SDC) databases. The proposed VAD is compared to the most com- monly used ones in the field in terms of speech/non-speech detection and recognition performance. Experimental results demonstrate a sustained advantage over G.729, AMR and AFE VADs.

This paper is mainly focused on showing experimental results of a feature extraction algorithm that combines spectral noise reduction and nonlinear feature normalization. The successfulness of this approach has been shown in a previous work, and in this one, we present several improvements that result in a performance comparable to that of the recently approved AFE for DSR. Noise reduction is now based on a Wiener filter instead of spectral subtraction. The voice activity detection based on the full-band energy has been replaced with a new one using spectral information. Relative improvements of 24.81% and 17.50% over our previous system are obtained for AURORA 2 and 3 respectively. Results for AURORA 2 are not as good as those for the AFE, but for AURORA 3 a relative improvement of 5.27% is obtained.

The need for both speed and increased precision in modern digital signal processing (DSP) applications represents a serious implementation obstacle. The paper explores the arithmetic benefits provided by the residue number system (RNS) for the design of such systems. Specifically, the fusion of the RNS with the popular distributed arithmetic (DA) is considered for the implementation of a discrete wavelet transform (DWT) filter bank. An exhaustive comparison of the advantages of RNS-DA over the traditional two's complement design, 2C-DA, is carried out for field-programmable logic (FPL), and cell-based ASIC technologies. The results show that the reported RNS-DA methodology, compared to a traditional 2C-DA design, enjoys a significant performance advantage that increases with precision.

A new methodology to compute the N-point DCT (Discrete Cosine Transform) in the QRNS (Quadratic Residue Number System) is presented, with a significant improvement in complexity and speed compared to the corresponding binary version. This reduction in the total number of arithmetic adders and multipliers is up to 21% and 26% for an 8-point and a 16-point DCT, respectively. In addition, large and slow binary adders and multipliers with a long carry propagation delay are replaced by high-speed small word-length modular adders and LUT (Look-Up Table) multipliers. When a Field Programmable Logic (FPL) implementation using Altera FLEX10KE devices of the proposed architecture for the 8-point DCT is considered, throughput is three times better than that obtained with the corresponding fixed point 2's complement binary implementation.