Conjugate mixture models for clustering multimodal data

The problem of multimodal clustering arises whenever the data are gathered with several physically different sensors. Observations from different modalities are not necessarily aligned in the sense there there is no obvious way to associate or compare them in some common space. A solution may consist in considering multiple clustering tasks independently for each modality. The main difficulty with such an approach is to guarantee that the unimodal clusterings are mutually consistent. In this letter, we show that multimodal clustering can be addressed within a novel framework: conjugate mixture models. These models exploit the explicit transformations that are often available between an unobserved parameter space (objects) and each of the observation spaces (sensors). We formulate the problem as a likelihood maximization task and derive the associated conjugate expectation-maximization algorithm. The convergence properties of the proposed algorithm are thoroughly investigated. Several local and global optimization techniques are proposed in order to increase its convergence speed. Two initialization strategies are proposed and compared. A consistent model selection criterion is proposed. The algorithm and its variants are tested and evaluated within the task of 3D localization of several speakers using both auditory and visual data.     

Towards multi-perspective rasterization

We present a novel framework for real-time multi-perspective rendering. While most existing approaches are based on ray-tracing, we present an alternative approach by emulating multi-perspective rasterization on the classical perspective graphics pipeline. To render a general multi-perspective camera, we first decompose the camera into piecewise linear primitive cameras called the general linear cameras or GLCs. We derive the closed-form projection equations for GLCs and show how to rasterize triangles onto GLCs via a two-pass rendering algorithm. In the first pass, we compute the GLC projection coefficients of each scene triangle using a vertex shader. The linear raster on the graphics hardware then interpolates these coefficients at each pixel. Finally, we use these interpolated coefficients to compute the projected pixel coordinates using a fragment shader. In the second pass, we move the pixels to their actual projected positions. To avoid holes, we treat neighboring pixels as triangles and re-render them onto the GLC image plane. We demonstrate our real-time multi-perspective rendering framework in a wide range of applications including synthesizing panoramic and omnidirectional views, rendering reflections on curved mirrors, and creating multi-perspective faux animations. Compared with the GPU-based ray tracing methods, our rasterization approach scales better with scene complexity and it can render scenes with a large number of triangles at interactive frame rates.     

Investigating the impact of video games on high school students’ engagement and learning about genetics

The popularity of video games has transcended entertainment crossing into the world of education. While the literature base on educational gaming is growing, there is still a lack of systematic study of this emerging technology’s efficacy. This quasi-experimental study evaluated a teacher created video game on genetics in terms of its affective and cognitive impact on student users. While statistical results indicated no differences (p > .05) in student learning as measured by our instrument, there were significant differences (p < .05) found in the participants’ level of engagement while interfacing with the video game. Implications on this emerging line of inquiry are discussed.     

Effects of combined wrist deviation and forearm rotation on discomfort score

The aim of the study was to examine the pattern of the change in discomfort for combined wrist deviation and forearm rotation as joint angles increased away from neutral in a repetitive task. There were five levels of wrist deviation (neutral, 35% and 55% of the range of motion (ROM) in radial and ulnar deviation) and five levels of forearm rotation (neutral, 30% and 60% of the ROM in pronation and supination). Twenty-five participants performed a repetitive flexion task with a force of 10 N +/- 1 N at a frequency of 15 exertions per min, with replication after 1 week for six of the participants. A visual analogue scale was used for recording the discomfort scores. Repeated measures analysis of covariance with the Greenhouse-Geisser correction, where necessary, was used on transformed values of the discomfort scores. Grip test endurance time at 50% of maximum voluntary contraction was included as a covariate. Wrist deviation (p = 0.007) and forearm rotation (p = 0.001) were found to have significant effects. Interactions of the main factors were not significant and nor was the covariate. Quadratic regression equations were derived and were used to generate iso-discomfort contours, which show a useful area of low discomfort around the central neutral zone of wrist postures, but with steep increases in discomfort at the extreme combinations of wrist ulnar/radial deviation with forearm pronation/supination. Discomfort equations and contours, showing wrist and forearm postures, which are either acceptable or potentially injurious, are useful for the design of industrial tools, machine controls and workspaces. Reference to these can help to reduce the risk of musculoskeletal injury associated with the tasks or tools by avoiding poor postures with unacceptable deviations from neutral posture.     

The between-day and inter-rater reliability of a novel wireless system to analyse lumbar spine posture

Lumbar posture is commonly assessed in non-specific chronic low back pain (NSCLBP), although quantitative measures have mostly been limited to laboratory environments. The BodyGuard? is a spinal position monitoring device that can monitor posture in real time, both inside and outside the laboratory. The reliability of this wireless device was examined in 18 healthy participants during usual sitting and forward bending, two tasks that are commonly provocative in NSCLBP. Reliability was determined using intraclass correlation coefficients (ICC), the standard error of measurement (SEM), the mean difference and the minimal detectable change (MDC90). Between-day ICC values ranged from 0.84 to 0.87, with small SEM (5%), mean difference (<9%) and MDC90 (<14%) values. Inter-rater ICC values ranged from 0.91 to 0.94, with small SEM (4%), mean difference (6%) and MDC90 (9%) values. Between-day and inter-rater reliability are essential requirements for clinical utility and were excellent in this study. Further studies into the validity of this device and its application in clinical trials in occupational settings are required. STATEMENT OF RELEVANCE: A novel device that can analyse spinal posture exposure in occupational settings in a minimally invasive manner has been developed. This study established that the device has excellent between-day and inter-rater reliability in healthy pain-free subjects. Further studies in people with low back pain are planned.     

Network trust management in emergency situations

We study the unique trust management, and more precisely reputation management and revocation of malicious nodes in the context of ad hoc networks used for emergency communications.Unlike in centralized systems, reputation management and revocation in ad hoc networks is non-trivial. This difficulty is due to the fact that the nodes have to collaboratively calculate the reputation value of a particular node and then revoke the node if the reputation value goes below a threshold. A major challenge in this scheme is to prevent a malicious node from discrediting other genuine nodes. The decision to revoke a node has to be communicated to all the nodes of the network. In traditional ad hoc networks the overhead of broadcasting the message throughout the network may be very high. We solve the problem of reputation management and node revocation in ad hoc networks of cell phones by using a threshold cryptography based scheme. Each node of the network would have a set of anonymous referees, which would store the reputation information of the node and issue reputation certificates to the node with timestamps. The misbehavior of a particular cell phone is reported to its anonymous referees, who issue certificates which reflect the positive and negative recommendations.     

Validation metrics for response histories: perspectives and case studies

The quantified comparison of transient response results are useful to analysts as they seek to improve and evaluate numerical models. Traditionally, comparisons of time histories on a graph have been used to make subjective engineering judgments as to how well the histories agree or disagree. Recently, there has been an interest in quantifying such comparisons with the intent of minimizing the subjectivity, while still maintaining a correlation with expert opinion. This increased interest has arisen from the evolving formalism of validation assessment where experimental and computational results are compared to assess computational model accuracy. The computable measures that quantify these comparisons are usually referred to as validation metrics. In the present work, two recently developed metrics are presented, and their wave form comparative quantification is demonstrated through application to analytical wave forms, measured and computed free-field velocity histories, and comparison with Subject Matter Expert opinion.

Kernel PCA for Feature Extraction and De-Noising in Nonlinear Regression

In this paper, we propose the application of the Kernel Principal Component Analysis (PCA) technique for feature selection in a high-dimensional feature space, where input variables are mapped by a Gaussian kernel. The extracted features are employed in the regression problems of chaotic Mackey–Glass time-series prediction in a noisy environment and estimating human signal detection performance from brain event-related potentials elicited by task relevant signals. We compared results obtained using either Kernel PCA or linear PCA as data preprocessing steps. On the human signal detection task, we report the superiority of Kernel PCA feature extraction over linear PCA. Similar to linear PCA, we demonstrate de-noising of the original data by the appropriate selection of various nonlinear principal components. The theoretical relation and experimental comparison of Kernel Principal Components Regression, Kernel Ridge Regression and ε-insensitive Support Vector Regression is also provided.     

A web-based e-learning system for increasing study efficiency by stimulating learner’s motivation

Due to the opportunities provided by the Internet, more and more people are taking advantage of distance learning courses and during the last few years enormous research efforts have been dedicated to the development of distance learning systems. So far, many e-learning systems are proposed and used practically. However, in these systems the e-learning completion rate is about 30%. One of the reasons is the low study desire when the learner studies the learning materials. In this research, we propose an interactive Web-based e-learning system. The purpose of our system is to increase the e-learning completion rate by stimulating learner’s motivation. The proposed system has three subsystems: the learning subsystem, learner support subsystem, and teacher support subsystem. The learning subsystem improves the learner’s study desire. The learner support subsystem supports the learner during the study, and the teacher support subsystem supports the teacher to get the learner’s study state. To evaluate the proposed system, we developed several experiments and surveys. By using new features such as: display of learner’s study history, change of interface color, encourage function, ranking function, self-determination of the study materials, and grouping of learners, the proposed system can increase the learning efficiency.