Andrés Marino Álvarez-Meza

Standardized evaluation of algorithms for computer-aided diagnosis of dementia based on structural MRI: The CADDementia challenge

Algorithms for computer-aided diagnosis of dementia based on structural MRI have demonstrated high performance in the literature, but are difficult to compare as different data sets and methodology were used for evaluation. In addition, it is unclear how the algorithms would perform on previously unseen data, and thus, how they would perform in clinical practice when there is no real opportunity to adapt the algorithm to the data at hand. To address these comparability, generalizability and clinical applicability issues, we organized a grand challenge that aimed to objectively compare algorithms based on a clinically representative multi-center data set. Using clinical practice as the starting point, the goal was to reproduce the clinical diagnosis. Therefore, we evaluated algorithms for multi-class classification of three diagnostic groups: patients with probable Alzheimers disease, patients with mild cognitive impairment and healthy controls. The diagnosis based on clinical criteria was used as reference standard, as it was the best available reference despite its known limitations. For evaluation, a previously unseen test set was used consisting of 354 T1-weighted MRI scans with the diagnoses blinded. Fifteen research teams participated with a total of 29 algorithms. The algorithms were trained on a small training set (n=30) and optionally on data from other sources (e.g., the Alzheimers Disease Neuroimaging Initiative, the Australian Imaging Biomarkers and Lifestyle flagship study of aging). The best performing algorithm yielded an accuracy of 63.0% and an area under the receiver-operating-characteristic curve (AUC) of 78.8%. In general, the best performances were achieved using feature extraction based on voxel-based morphometry or a combination of features that included volume, cortical thickness, shape and intensity. The challenge is open for new submissions via the web-based framework: http://caddementia.grand-challenge.org.

Global and local choice of the number of nearest neighbors in locally linear embedding

Highlights? We propose a new method for automatically computing the number of neighbors in LLE. ? We analyze the global and local properties of the embedding results. ? We study manifolds where the density and the intrinsic dimensionality of the neighborhoods are variable. ? Artificial and real-world datasets were tested. The crux in the locally linear embedding algorithm is the selection of the number of nearest neighbors k. Some previous techniques have been developed for finding this parameter based on embedding quality measures. Nevertheless, they do not achieve suitable results when they are tested on several kind of manifolds. In this work is presented a new method for automatically computing the number of neighbors by means of analyzing global and local properties of the embedding results. Besides, it is also proposed a second strategy for choosing the parameter k, on manifolds where the density and the intrinsic dimensionality of the neighborhoods are changeful. The first proposed technique, called preservation neighborhood error, calculates a unique value of k for the whole manifold. Moreover, the second method, named local neighborhood selection, computes a suitable number of neighbors for each sample point in the manifold. The methodologies were tested on artificial and real-world datasets which allow us to visually confirm the quality of the embedding. According to the results our methods aim to find suitable values of k and appropriated embeddings.

Time-series discrimination using feature relevance analysis in motor imagery classification

Abstract The use of motor imagery discrimination using feature relevance analysis (MIDFR) is investigated for classification tasks based on electroencephalography (EEG) signals. The method addresses the problem of a direct and automatic finding of the time-varying features influencing the most on distinguishing motor imagery tasks. The method introduces a stochastic relevance stage that is primarily used for properly handling the set of short-time features, which are extracted as to make prominent the nonstationary behavior of the EEG data. Furthermore, since it is widely accepted that the motor imagery information is concentrated in the μ and β neural activity bands, we make use of the empirical mode decomposition together with the common-spatial-patterns mapping. We test two different motor imagery databases, using a soft-margin support vector machine classifier that is validated by a 10-fold cross-validation methodology. Since the proposed MIDFR algorithm better encodes neural activity dynamics, experimental results carried out, which are also contrasted with other state-of-the art approaches, show that the proposed approach allows improving detection of MI classification tasks. Besides, the computed relevance on the EEG channels is in accordance with other clinical findings reported in the literature.

Kernel based hand gesture recognition using kinect sensor

Category 4. A machine learning based methodology is proposed to recognize a predefined set of hand gestures using depth images. For such purpose, a RGBD sensor (Microsoft kinect) is employed to track the hand position. Thus, a preprocessing stage is presented to subtract the region of interest from depth images. Moreover, a learning algorithm based on kernel methods is used to discover the relationships among samples, properly describing the studied gestures. Proposed methodology aims to obtain a representation space which allow us to identify the dynamic of hand movements. Attained results show how our approach presents a suitable performance for detecting different hand gestures. As future work, we are interested in recognize more complex human activities, in order to support the development of human-computer interface systems.

Image Segmentation Based on Multi-Kernel Learning and Feature Relevance Analysis

In this paper an automatic image segmentation methodology based on Multiple Kernel Learning (MKL) is proposed. In this regard, we compute some image features for each input pixel, and then combine such features by means of a MKL framework. We automatically fix the weights of the MKL approach based on a relevance analysis over the original input feature space. Moreover, an unsupervised image segmentation measure is used as a tool to establish the employed kernel free parameter. A Kernel Kmeans algorithm is used as spectral clustering method to segment a given image. Experiments are carried out aiming to test the efficiency of the incorporation of weighted feature information into clustering procedure, and to compare the performance against state of the art algorithms, using a supervised image segmentation measure. Attained results show that our approach is able to compute a meaningful segmentations, demonstrating its capability to support further vision computer applications.

Unsupervised Kernel Function Building Using Maximization of Information Potential Variability

We propose a kernel function estimation strategy to support machine learning tasks by analyzing the input samples using Renyi’s Information Metrics. Specifically, we aim to identify a Reproducing Kernel Hilbert Space spanning the most widely the information force among data points by the maximization of the information potential variability of Parzen-based pdf estimation. So, a Gaussian kernel bandwidth updating rule is obtained as a function of the forces induced by a given dataset. Our proposal is tested on synthetic and real-world datasets related to clustering and classification tasks. Obtained results show that presented approach allows to compute RKHS’s favoring data groups separability, attaining suitable learning performances in comparison with state of the art algorithms.

Motor Imagery Classification for BCI Using Common Spatial Patterns and Feature Relevance Analysis

Recently, there have been many efforts to develop Brain Computer Interface (BCI) systems, allowing to identify and discriminate brain activity. In this work, a Motor Imagery (MI) discrimination framework is proposed, which employs Common Spatial Patterns (CSP) as preprocessing stage, and a feature relevance analysis approach based on an eigendecomposition method to identify the main features that allow to discriminate the studied EEG signals. The CSP is employed to reveal the dynamics of interest from EEG signals, and then we select a set of features representing the best as possible the studied process. EEG signals modeling is done by feature estimation of three frequency-based and one time-based. Besides, a relevance analysis over the EEG channels is performed, which gives to the user an idea about the channels that mainly contribute for the MI discrimination. Our approach is tested over a well known MI dataset. Attained results (95.21±4.21 [%] mean accuracy) show that presented framework can be used as a tool to support the discrimination of MI brain activity.

Feature relevance analysis supporting automatic motor imagery discrimination in EEG based BCI systems

Recently, there have been many efforts to develop Brain Computer Interface (BCI) systems, allowing identifying and discriminating brain activity, as well as, support the control of external devices, and to understand cognitive behaviors. In this work, a feature relevance analysis approach based on an eigen decomposition method is proposed to support automatic Motor Imagery (MI) discrimination in electroencephalography signals for BCI systems. We select a set of features representing the best as possible the studied process. For such purpose, a variability study is performed based on traditional Principal Component Analysis. EEG signals modelling is carried out by feature estimation of three frequency-based and one time-based. Our approach provides testing over a well-known MI dataset. Attained results show that presented algorithm can be used as tool to support discrimination of MI brain activity, obtaining acceptable results in comparison to state of the art approaches.

Video Segmentation Framework Based on Multi-kernel Representations and Feature Relevance Analysis for Object Classification

A video segmentation framework to automatically detect moving objects in a scene using static cameras is proposed. Using Multiple Kernel Representations, we aim to enhance the data separability into the scene by incorporating multiple information sources into the process, and employing a relevance analysis each source is automatically weighted. A tuned Kmeans technique is employed to group pixels as static or moving objects. Moreover, the proposed methodology is tested for the classification of people and abandoned objects. Attained results over real-world datasets, show how our approach is stable using the same parameters for all experiments.

Motor imagery classification using feature relevance analysis: An Emotiv-based BCI system

Brain Computer Interfaces (BCI) have been emerged as an alternative to support automatic systems able to interpret brain functions, commonly, by analyzing electroencephalography (EEG) recordings. In this work, a time-series discrimination methodology, called Motor Imagery Discrimination by Relevance Analysis (MIDRA), is presented to support the development of BCI from EEG data. Particularly, a Motor Imagery (MI) paradigm is studied, i.e., imagination of left-right hand movements. In this sense, a feature relevance analysis strategy is presented to select representing characteristics using a variability criterion. Besides, short-time parameters are estimated from EEG data by considering both time and time-frequency representations to deal with non-stationary dynamics. MIDRA is assessed on two different BCI databases, a well-known MI data and an Emotiv-based dataset. Attained results showed that MIDRA enhances the BCI system performance in comparison with benchmark methods by suitable ranking the input feature set. Moreover, applying MIDRA in a BCI based on the Emotiv device is a straightforward alternative for dealing with MI paradigms.