Expectation Conditional Maximization-Based Deformable Shape Registration

This paper addresses the issue of matching statistical and non-rigid shapes, and introduces an Expectation Conditional Maximization-based deformable shape registration (ECM-DSR) algorithm. Similar to previous works, we cast the statistical and non-rigid shape registration problem into a missing data framework and handle the unknown correspondences with Gaussian Mixture Models (GMM). The registration problem is then solved by fitting the GMM centroids to the data. But unlike previous works where equal isotropic covariances are used, our new algorithm uses heteroscedastic covariances whose values are iteratively estimated from the data. A previously introduced virtual observation concept is adopted here to simplify the estimation of the registration parameters. Based on this concept, we derive closed-form solutions to estimate parameters for statistical or non-rigid shape registrations in each iteration. Our experiments conducted on synthesized and real data demonstrate that the ECM-DSR algorithm has various advantages over existing algorithms.

Fully Automatic Segmentation of AP Pelvis X-rays via Random Forest Regression and Hierarchical Sparse Shape Composition

Knowledge of landmarks and contours in anteroposterior (AP) pelvis X-rays is invaluable for computer aided diagnosis, hip surgery planning and image-guided interventions. This paper presents a fully automatic and robust approach for landmarking and segmentation of both pelvis and femur in a conventional AP X-ray. Our approach is based on random forest regression and hierarchical sparse shape composition. Experiments conducted on 436 clinical AP pelvis x-rays show that our approach achieves an average point-to-curve error around 1.3 mm for femur and 2.2 mm for pelvis, both with success rates around 98%. Compared to existing methods, our approach exhibits better performance in both the robustness and the accuracy.

Sparse representation based action and gesture recognition

In this paper we present a solution to the problem of action and gesture recognition using sparse representations. The dictionary is modelled as a simple concatenation of features computed for each action or gesture class from the training data, and test data is classified by finding sparse representation of the test video features over this dictionary. Our method does not impose any explicit training procedure on the dictionary. We experiment our model with two kinds of features, by projecting (i) Gait Energy Images (GEIs) and (ii) Motion-descriptors, to a lower dimension using Random projection. Experiments have shown 100% recognition rate on standard datasets and are compared to the results obtained with widely used SVM classifier.

A Closed-Form, Consistent and Robust Solution to Uncalibrated Photometric Stereo Via Local Diffuse Reflectance Maxima

Images of an object under different illumination are known to provide strong cues about the object surface. A mathematical formalization of how to recover the normal map of such a surface leads to the so-called uncalibrated photometric stereo problem. In the simplest instance, this problem can be reduced to the task of identifying only three parameters: the so-called generalized bas-relief (GBR) ambiguity. The challenge is to find additional general assumptions about the object, that identify these parameters uniquely. Current approaches are not consistent, i.e., they provide different solutions when run multiple times on the same data. To address this limitation, we propose exploiting local diffuse reflectance (LDR) maxima, i.e., points in the scene where the normal vector is parallel to the illumination direction (see Fig. 1). We demonstrate several noteworthy properties of these maxima: a closed-form solution, computational efficiency and GBR consistency. An LDR maximum yields a simple closed-form solution corresponding to a semi-circle in the GBR parameters space (see Fig. 2); because as few as two diffuse maxima in different images identify a unique solution, the identification of the GBR parameters can be achieved very efficiently; finally, the algorithm is consistent as it always returns the same solution given the same data. Our algorithm is also remarkably robust: It can obtain an accurate estimate of the GBR parameters even with extremely high levels of outliers in the detected maxima (up to 80 % of the observations). The method is validated on real data and achieves state-of-the-art results.

Low rank subspace clustering (LRSC)

We consider the problem of fitting a union of subspaces to a collection of data points drawn from one or more subspaces and corrupted by noise and/or gross errors. We pose this problem as a non-convex optimization problem, where the goal is to decompose the corrupted data matrix as the sum of a clean and self-expressive dictionary plus a matrix of noise and/or gross errors. By self-expressive we mean a dictionary whose atoms can be expressed as linear combinations of themselves with low-rank coefficients. In the case of noisy data, our key contribution is to show that this non-convex matrix decomposition problem can be solved in closed form from the SVD of the noisy data matrix. The solution involves a novel polynomial thresholding operator on the singular values of the data matrix, which requires minimal shrinkage. For one subspace, a particular case of our framework leads to classical PCA, which requires no shrinkage. For multiple subspaces, the low-rank coefficients obtained by our framework can be used to construct a data affinity matrix from which the clustering of the data according to the subspaces can be obtained by spectral clustering. In the case of data corrupted by gross errors, we solve the problem using an alternating minimization approach, which combines our polynomial thresholding operator with the more traditional shrinkage-thresholding operator. Experiments on motion segmentation and face clustering show that our framework performs on par with state-of-the-art techniques at a reduced computational cost.

Parsing for agile modeling

In order to analyze software systems, it is necessary to model them. Static software models are commonly imported by parsing source code and related data. Unfortunately, building custom parsers for most programming languages is a non-trivial endeavour. This poses a major bottleneck for analyzing software systems programmed in languages for which importers do not already exist. Luckily, initial software models do not require detailed parsers, so it is possible to start analysis with a coarse-grained importer, which is then gradually refined. In this paper we propose an approach to "agile modeling" that exploits island grammars to extract initial coarse-grained models, parser combinators to enable gradual refinement of model importers, and various heuristics to recognize language structure, keywords and other language artifacts.

Online social networks: why we disclose

On online social networks such as Facebook, massive self-disclosure by users has attracted the attention of industry players and policymakers worldwide. Despite the impressive scope of this phenomenon, very little is understood about what motivates users to disclose personal information. Integrating focus group results into a theoretical privacy calculus framework, we develop and empirically test a Structural Equation Model of self-disclosure with 259 subjects. We find that users are primarily motivated to disclose information because of the convenience of maintaining and developing relationships and platform enjoyment. Countervailing these benefits, privacy risks represent a critical barrier to information disclosure. However, users’ perception of risk can be mitigated by their trust in the network provider and availability of control options. Based on these findings, we offer recommendations for network providers.

"Deal of the day” platforms: what drives consumer loyalty?

Maintaining a loyal customer base is challenging for “Deal of the Day” (DoD) platforms. DoD providers market and sell deals on products and services, yet it is the merchants who ultimately deliver those to consumers. Low entry and switching costs drive competition in this market. However, research on the determinants of user loyalty in the DoD context is limited. This study uses Grounded Theory and Structural Equation Modeling to explore the phenomenon of DoD platform loyalty. Particularly, monetary benefits, signal-to-noise ratio, perceived risk, and service friendliness during a merchant encounter emerge as powerful determinants of loyalty in this novel context.

Self-disclosure and privacy calculus on social networking sites: the role of culture

Social Networking Sites (SNSs) have become extremely popular around the world. They rely on user-generated content to offer engaging experience to its members. Cultural differences may influence the motivation of users to create and share content on SNS. This study adopts the privacy calculus perspective to examine the role of culture in individual self-disclosure decisions. The authors use structural equation modeling and multi-group analysis to investigate this dynamics. The findings reveal the importance of cultural dimensions of individualism and uncertainty avoidance in the cognitive processes of SNS users.

Panel report ECIS 2012: publication strategy for junior researchers: quantity vs. quality, importance of the first authorship and collaboration

The publication record is a key component of a successful academic career in IS. Despite its importance, its definition - especially for junior researchers―remains unclear. Is it better to have one A-publication or three Bpublications? Does being the third author on an A-publication carry more weight than being the first author on a Bpublication? Is it better to publish with as few co-authors as possible to demonstrate ability for independent work or is publishing with others a sign of good teamwork and academic excellence? Faced with these uncertainties, young researchers increasingly question the choices they make regarding their publication strategy. If unaddressed, these issues are bound to interfere with the quality of the IS research and scholars’ job satisfaction. This article raises these concerns associated with a publication strategy for junior researchers and reports the views voiced by five academics at a panel session at the European Conference on Information Systems 2012. In particular, the following topics are discussed: quantity vs. quality, value of the first authorship, the “optimal” number of authors, and the issues of co-authorship with an academic supervisor.

Privacy concerns and identity in online social networks

Driven by privacy-related fears, users of Online Social Networks may start to reduce their network activities. This trend can have a negative impact on network sustainability and its business value. Nevertheless, very little is understood about the privacy-related concerns of users and the impact of those concerns on identity performance. To close this gap, we take a systematic view of user privacy concerns on such platforms. Based on insights from focus groups and an empirical study with 210 subjects, we find that (i) Organizational Threats and (ii) Social Threats stemming from the user environment constitute two underlying dimensions of the construct “Privacy Concerns in Online Social Networks”. Using a Structural Equation Model, we examine the impact of the identified dimensions of concern on the Amount, Honesty, and Conscious Control of individual self-disclosure on these sites. We find that users tend to reduce the Amount of information disclosed as a response to their concerns regarding Organizational Threats. Additionally, users become more conscious about the information they reveal as a result of Social Threats. Network providers may want to develop specific mechanisms to alleviate identified user concerns and thereby ensure network sustainability.