Time filter

Source Type

Ram J.,University of Adelaide | Corkindale D.,University of Western Australia | Wu M.-L.,United International College
Journal of Computer Information Systems | Year: 2013

This study identifies the key antecedent factors for accomplishing the adoption stage of enterprise resource planning (ERP) systems. Five potential antecedent factors of adoption were derived from the literature, including that on innovation theories, and data were obtained from a sample of 217 organizations across Australia. A structural equation modeling (SEM) technique was used to examine the complex relationships between antecedents and the adoption decision. We found that there were three positive drivers of a successful outcome of the ERP adoption stage. Prior findings have shown that system quality is a key enabler for innovation adoption by individuals, and we found that system quality is also an important driver for organizational adoption of ERP. It was also indicated that organizations consider adopting ERP when the market and customer patterns are relatively stable rather than in turbulent environments.

Shao L.,University of Sheffield | Gao R.,Leiden University | Liu Y.,Hong Kong Polytechnic University | Zhang H.,United International College | Zhang H.,PKU HKUST Shenzhen Hong Kong Institution
Neurocomputing | Year: 2011

Classic transformation methods have been widely and efficiently used in image processing areas, such as image de-noising, image segmentation, feature detection, and compression. Based on their compact signal and image representation ability, we apply the transform based techniques on the video recognition area to extract discriminative information from each given video sequence, and use the transformed coefficients as descriptors for representing and recognizing human actions in video sequences. We validate our proposed methods on the KTH and the Hollywood datasets, which have been extensively studied by a lot of researchers. The proposed descriptors, especially the wavelet transform based descriptor, yield promising results on action recognition. © 2010 Elsevier B.V.

Xia Y.,Tsinghua University | Su W.,United International College | Lau R.Y.K.,City University of Hong Kong | Liu Y.,Tsinghua University
Enterprise Information Systems | Year: 2013

Unlike most online social networks where explicit links among individual users are defined, the relations among commercial entities (e.g. firms) may not be explicitly declared in commercial Web sites. One main contribution of this article is the development of a novel computational model for the discovery of the latent relations among commercial entities from online financial news. More specifically, a CRF model which can exploit both structural and contextual features is applied to commercial entity recognition. In addition, a point-wise mutual information (PMI)-based unsupervised learning method is developed for commercial relation identification. To evaluate the effectiveness of the proposed computational methods, a prototype system called CoNet has been developed. Based on the financial news articles crawled from Google finance, the CoNet system achieves average F-scores of 0.681 and 0.754 in commercial entity recognition and commercial relation identification, respectively. Our experimental results confirm that the proposed shallow natural language processing methods are effective for the discovery of latent commercial networks from online financial news. © 2013 Copyright Taylor and Francis Group, LLC.

Li X.,Hong Kong Baptist University | Mo L.,United International College | Yuan X.,Hong Kong Baptist University | Zhang J.,United International College
Computational Statistics and Data Analysis | Year: 2014

The least absolute shrinkage and selection operator (LASSO) has been playing an important role in variable selection and dimensionality reduction for linear regression. In this paper we focus on two general LASSO models: Sparse Group LASSO and Fused LASSO, and apply the linearized alternating direction method of multipliers (LADMM for short) to solve them. The LADMM approach is shown to be a very simple and efficient approach to numerically solve these general LASSO models. We compare it with some benchmark approaches on both synthetic and real datasets. © 2014 Elsevier B.V. All rights reserved.

Ma A.J.,Hong Kong Baptist University | Yuen P.C.,Hong Kong Baptist University | Yuen P.C.,United International College | Li J.,Hong Kong Baptist University
Proceedings of the IEEE International Conference on Computer Vision | Year: 2013

This paper addresses a new person re-identification problem without the label information of persons under non-overlapping target cameras. Given the matched (positive) and unmatched (negative) image pairs from source domain cameras, as well as unmatched (negative) image pairs which can be easily generated from target domain cameras, we propose a Domain Transfer Ranked Support Vector Machines (DTRSVM) method for re-identification under target domain cameras. To overcome the problems introduced due to the absence of matched (positive) image pairs in target domain, we relax the discriminative constraint to a necessary condition only relying on the positive mean in target domain. By estimating the target positive mean using source and target domain data, a new discriminative model with high confidence in target positive mean and low confidence in target negative image pairs is developed. Since the necessary condition may not truly preserve the discriminability, multi-task support vector ranking is proposed to incorporate the training data from source domain with label information. Experimental results show that the proposed DTRSVM outperforms existing methods without using label information in target cameras. And the top 30 rank accuracy can be improved by the proposed method upto 9.40% on publicly available person re-identification datasets. © 2013 IEEE.

Huang H.,Hong Kong Baptist University | Huang H.,United International College | Zhang H.,United International College | Cheung Y.-M.,Hong Kong Baptist University | Cheung Y.-M.,United International College
Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition | Year: 2016

How to avoid ambiguity is a challenging problem for conic-based homography estimation. In this paper, we address the problem of homography estimation from two separate ellipses. We find that any two ellipses have a unique common self-polar triangle, which can provide three line correspondences. Furthermore, by investigating the location features of the common self-polar triangle, we show that one vertex of the triangle lies outside of both ellipses, while the other two vertices lies inside the ellipses separately. Accordingly, one more line correspondence can be obtained from the intersections of the conics and the common self-polar triangle. Therefore, four line correspondences can be obtained based on the common self-polar triangle, which can provide enough constraints for the homography estimation. The main contributions in this paper include: (1) A new discovery on the location features of the common self-polar triangle of separate ellipses. (2) A novel approach for homography estimation. Simulate experiments and real experiments are conducted to demonstrate the feasibility and accuracy of our approach.

Jia H.,Hong Kong Baptist University | Cheung Y.-M.,Hong Kong Baptist University | Cheung Y.-M.,United International College | Liu J.,Hong Kong Baptist University
Pattern Recognition | Year: 2014

Competitive learning approaches with individual penalization or cooperation mechanisms have the attractive ability of automatic cluster number selection in unsupervised data clustering. In this paper, we further study these two mechanisms and propose a novel learning algorithm called Cooperative and Penalized Competitive Learning (CPCL), which implements the cooperation and penalization mechanisms simultaneously in a single competitive learning process. The integration of these two different kinds of competition mechanisms enables the CPCL to locate the cluster centers more quickly and be insensitive to the number of seed points and their initial positions. Additionally, to handle nonlinearly separable clusters, we further introduce the proposed competition mechanism into kernel clustering framework. Correspondingly, a new kernel-based competitive learning algorithm which can conduct nonlinear partition without knowing the true cluster number is presented. The promising experimental results on real data sets demonstrate the superiority of the proposed methods. © 2014 Elsevier Ltd.

Feng X.,United International College | Parnas D.L.,University of Limerick | Tse T.H.,University of Hong Kong
IEEE Transactions on Software Engineering | Year: 2011

Tabular expressions have been proposed as a notation to document mathematically precise but readable software specifications. One of the many roles of such documentation is to guide testers. This paper 1) explores the application of four testing strategies (the partition strategy, decision table-based testing, the basic meaningful impact strategy, and fault-based testing) to tabular expression-based specifications, and 2) compares the strategies on a mathematical basis through formal and precise definitions of the subsumption relationship. We also compare these strategies through experimental studies. These results will help researchers improve current methods and will enable testers to select appropriate testing strategies for tabular expression-based specifications. © 2011 IEEE.

Chen Q.,United International College | Zhang H.,United International College
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2015

This paper addresses the problem of motion and shape recovery from a two-mirror system which is able to generate five views of an object. Different from existing methods, this paper uses a short video instead of static snapshots so that it can help with action recognition once the 3D visual hull model is reconstructed. In order to solve the problem, this paper shows the geometry relationship between the two-mirror system and circular motion, so that the two-mirror system can be solved as circular motions. Different from the approach of Zhang et al. [22], we avoid using the vanishing point of X-axis which would cause accumulate error when calculating the epipoles of two views. Results of comparative experiments and the 3D visual hull of model show the feasibility and the accuracy of the proposed approach. © Springer International Publishing Switzerland 2015.

Cheung D.,United International College | Cucker F.,City University of Hong Kong
SIAM Journal on Optimization | Year: 2013

We give an O(log n) bound for the expectation of the logarithm of the condition number K(A, b, c) introduced in "Solving Linear Programs with Finite Precision: I. Condition Numbers and Random Programs" [Math. Program., 99 (2004), pp. 175-196]. This bound improves the previously existing bound, which was of O(n), and yields average-case bounds for both the required precision and the complexity of computing an optimal basis (or a pair of primal-dual optimizers). © 2013 Society for Industrial and Applied Mathematics.

Loading United International College collaborators
Loading United International College collaborators