Previous work has demonstrated that the image variations of many objects (human faces in particular) under variable lighting can be effectively modeled by low dimensional linear spaces. Basis images spanning this space are usually obtained one two ways: A large number object different conditions is acquired, and principal component analysis (PCA) used to estimate a subspace. Alternatively, 3D model (perhaps reconstructed from images) render virtual either point sources which subspace derived...

We present a new Gaussian process (GP) inference algorithm, called online sparse matrix processes (OSMGP), and demonstrate its merits by applying it to the problems of head pose estimation visual tracking. The OSMGP is based upon observation that for kernels with local support, Gram typically sparse. Maintaining updating Cholesky factor can be done efficiently using Givens rotations. This leads an exact, algorithm whose update time scales linearly size matrix. Further, we provide method...

We propose an algorithm for accurate tracking of (articulated) objects using online update appearance and shape. The challenge here is to model foreground with histograms in a way that both efficient accurate. In this algorithm, the constantly changing shape modeled as small number rectangular blocks, whose positions within window are adaptively determined. Under general assumption stationary appearance, we show robust object possible by adjusting locations these blocks. Implemented MATLAB...

This study explores the effects of perspective-taking non-player characters (NPCs) on enhancing game immersion in prosocial virtual reality (VR) games. Prosocial games are focusing helping others. Game researchers have been keen to investigate factors that influence immersive experience digital Previous studies show VR allows people take perspective others, inducing empathy and behaviour real world. In this lab-based study, we explore whether how taking other – NPCs a influences players'...

This paper proposes a dictionary learning framework that combines the proposed block/group (BGSC) or reconstructed (R-BGSC) sparse coding schemes with novel Intra-block Coherence Suppression Dictionary Learning algorithm. An important and distinguishing feature of is all blocks are trained simultaneously respect to each data group while intra-block coherence being explicitly minimized as an objective. We provide both empirical evidence heuristic support for this can be considered direct...

This paper proposes a new affine registration algorithm for matching two point sets in IR <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> or xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> . The input are represented as probability density functions, using either Gaussian mixture models discrete models, and the problem of registering is treated aligning distributions. Since polynomials transform symmetric tensors under an...

This paper aims to bridge the affective gap between image content and emotional response of viewer it elicits by using High-Level Concepts (HLCs). In contrast previous work that relied solely on low-level features or used convolutional neural network (CNN) as a blackbox, we use HLCs generated pretrained CNNs in an explicit way investigate relations/associations these (small) set Ekman's classes. As proof-of-concept, first propose linear admixture model for modeling relations, resulting...

In clinical applications where structural asymmetries between homologous shapes have been correlated with pathology, the questions of definition and quantification "asymmetry" arise naturally. When not only degree but position deformity is thought relevant, asymmetry localization must also be addressed. Asymmetries paired already formulated in terms (nonrigid) diffeomorphisms shapes. For infinity such maps possible for a given pair, we define optimality as minimization deviation from...

We propose a novel affine registration algorithm for matching 2D feature points. Unlike many previously published work on point matching, the proposed does not require any optimization and in absence of data noise, will recover exact transformation unknown correspondence. The two-step first reduces general case to orthogonal case, rotation is computed as roots low-degree polynomial with complex coefficients. algebraic geometric ideas behind method are both clear transparent, its...

This paper proposes a novel image-retargeting algorithm that can retarget images to large family of non-rectangular shapes. Specifically, we study image retargeting from broader perspective includes the content as well shape an image, and proposed shape-aware (CASAIR) is driven by dual objectives preservation domain transformation, with latter defined application-specific target shape. The based on idea seam segment carving successively removes low-cost segments simultaneously achieve two...

SIGNIFICANCE A majority of visually impaired older adults were able to learn proficiently use visual-assistive iPhone applications (apps) following a median 1 hour and/or multiple training sessions, which should be considered when planning vision rehabilitation service delivery, including the option for remote telerehabilitation those who prefer that modality. PURPOSE Older with low are increasingly using technology improve their visual functioning. We examined whether age-related...

While general self-efficacy is known to relate achievement in many areas, it has rarely been evaluated individuals with low vision. Here we explore the psychometric properties and targeting of New General Self Efficacy Scale (NGSES) using Rasch analysis data collected from older adult clinical trial participants Participants (n = 121) completed pre-intervention telephone questionnaires (i.e., NGSES, Activity Inventory (AI), Beck Depression (BDI), SF-36, Telephone Interview for Cognitive...

Symmetric positive-definite (SPD) matrices are ubiquitous in Computer Vision, Machine Learning and Medical Image Analysis. Finding the center/average of a population such is common theme many algorithms as clustering, segmentation, principal geodesic analysis, etc. The center can be defined using variety distance/divergence measures minimizer sum squared distances/divergences from unknown to members population. It well known that computation Karcher mean for space SPD which negatively-curved...

This paper proposes a novel approach for sparse coding that further improves upon the representation-based classification (SRC) framework. The proposed framework, Affine-Constrained Group Sparse Coding (ACGSC), extends current SRC framework to problems with multiple input samples. Geometrically, affineconstrained group essentially searches vector in convex hull spanned by vectors can best be coded using given dictionary. resulting objective function is still and efficiently optimized...

In this paper, we propose a novel framework for computing single or multiple atlases (templates) from large population of images. Unlike many existing methods, our proposed approach is distinguished by its emphasis on the sharpness computed and requirement rotational invariance. particular, argue that sharp atlas images retain crucial important anatomical features with high fidelity are more useful medical imaging applications when compared blurry fuzzy most methods. The geometric notion...

This paper reports on a text mining-based case study aimed at determining how virtual reality (VR) games, as examples of really new products (RNPs), market themselves when they are introduced to the mass market. The goal was examine marketing foci RNPs and any subsequent changes over time survive. VR games type RNP offering several unique benefits, such immersive gameplay storytelling, which advanced compared with their earlier counterparts. To we collected 17,000 pieces promotional from...

As an interdisciplinary field, data-driven journalism integrates the intellectual origins of investigative journalism, computer-assisted reporting, and emerging paradigm computational social science. Studies news production have revealed, however, that professionals are reinforcing existing power structures via interpretive community, where homophily-evoked interactions—even in media context—create echo chambers discussion fragmentation. Is representation electronic public sphere breaking...

We present an algorithm that uses partitioning and gluing to compress large triangular meshes which are too complex fit in main memory. The is based largely on the existing mesh compression algorithms, most of require 'in-core' representation input mesh. Our solution partition into smaller submeshes these separately using techniques. Since a direct out question, instead, we simplified use model obtain original model. In order recover full connectivity, simple scheme for encoding/decoding...