Guangdong University of Education

China

Guangdong University of Education

China
SEARCH FILTERS
Time filter
Source Type

Zhang S.,Guangzhou University | Li J.,Guangzhou University | Huang H.,Guangdong University of Education | Wu H.,Guangzhou University | And 2 more authors.
Proceedings - 2016 9th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics, CISP-BMEI 2016 | Year: 2016

Analysis of course survey data are foundationally important because it has a close relation with students' skills training and talent development. However, there are multiple categories of classification information that characterized different students from different perspectives, such as genders, classes, types or regions. Compared to the correlation between survey data with individual kind of classification, consensus between those with multiple categories is more informative. However, traditional methods can not handle this problem with multiple kinds of classification categories. In view of the problem above, in this work, we proposed to conduct a new factor analysis method in the consensus fashion. Specifically, we computed a consensus structure of different categories of classifications and the similarity matrices for each factor, respectively. Motivated by the famous clustering measure RAND, we proposed a generalized method called Rmm which is used to compute the consistency of two different similarity matrices, with the option to adopt weights. Followed the similar spirit of Rmm, we also computed pairwise consistency between different factors without category information. Experimental results show that: (i) Our Rmm method is capable to combine different categories of classification information in an effective manner;(ii) our Rmm method preserve promising linear property for the combination of classifications; and (iii) Our Rmm method can be used in the case of fuzzy similarity, which is more applicable in different scenarios. With the proposed method, this work attempted to figure out the veiled laws and covered from read data, with the hope to provide suggestions for course arrangement and personnel cultivating program in universities, and to shed light on personal development of university students who major on computer science. © 2016 IEEE.


Zhu X.,Guangdong University of Education | Li H.,Guangdong University of Education | Shi Z.,Guangdong University of Technology | Xiang Y.,Guangdong University of Technology | He Y.,Guangdong Polytechnic Normal University
Journal of Physics B: Atomic, Molecular and Optical Physics | Year: 2017

We investigate the properties of gap solitons in spin-orbit-coupled Bose-Einstein condensates in mixed linear-nonlinear optical lattices. The mixed linear-nonlinear optical lattices can support parity-time (PT)-symmetric soliton solutions, and these PT-symmetric solitons can stably exist in the semi-infinite gap of linear optical lattices. The PT-symmetric gap-stripe solitons are found by increasing the strength of the spin-orbit coupling. It is found that the amplitude of the nonlinear optical lattices can significantly affect the stability of these gap and gap-stripe solitons. With an increase of the amplitude of the nonlinear optical lattices, the stable domains of the gap and gap-stripe solitons are shrunk. The unstable solitons can show unique evolution characteristics and the stability analyses are also confirmed by the evolution simulations. © 2017 IOP Publishing Ltd.


Chen G.,Guangdong University of Education | Chen Q.,Guangdong University of Education | Zhang D.,Sun Yat Sen University
Proceedings - 2014 International Conference on Digital Home, ICDH 2014 | Year: 2014

In this paper we proposed a dictionary learning and dimensionality reduction (DLDR) scheme for image steganalysis. We construct a structural discriminative dictionary which is learned from the reduced dimension space and exploit the discriminative information in stego-images. Simulation results verify the effectiveness of the proposed approach and the performance is considerable. Both the dictionary and sparse coding can be correctly computed and the learned dictionary using the proposed method can be used to improve image steganlysis. © 2014 IEEE.


Chen G.,Guangdong University of Education | Chen Q.,Guangdong University of Education | Zhang D.,Sun Yat Sen University
2015 12th International Conference on Fuzzy Systems and Knowledge Discovery, FSKD 2015 | Year: 2015

Image steganalysis based on supervised distance metric learning is to find an appropriate measure of similarity between image features where the distribution discrepancy between cover-images and stego-images are analyzed in the reduced dimensional space. Our approach is novel in that it combines the merits of weight metric learning and image distribution analysis in reduced dimension space. By this learning metrics, we exploit a new steganalysis metric to discriminate stego-images from clean images. The experiment results show the effectiveness of the propose approach for some data hiding method. © 2015 IEEE.


Chen G.,Guangdong University of Education | Chen Q.,Guangdong University of Education | Zhang D.,Sun Yat Sen University
Proceedings - 2015 8th International Congress on Image and Signal Processing, CISP 2015 | Year: 2015

To investigate the presence of hidden information in cover photographic images is very important for image steganalysis at the present time. Steganalysis can be also regarded as a pattern recognition classification problem to decide which class a test image is classified as: the innocent photographic image or the stego-image. In this paper we propose an Randomized Neural Network (RNN), based multi-modality classifier to improve the accuracy of image steganalysis. In this work: multi-modality steganalysis may provide complementary information to discriminate stego-images from innocent images. Experiments results show that our multimodal scheme can effectively promote the accuracy of image steganalysis and achieve performance at high speed. We also achieve a classification accuracy of 93.43% when combining all five modalities of steganalysis model, and only 91.33% when using even the best individual modality of steganalysis model. © 2015 IEEE.


Chen G.,Guangdong University of Education | Chen Q.,Guangdong University of Education | Zhang D.,Sun Yat Sen University | Chen Y.,Guangdong University of Education
Journal of Computational Information Systems | Year: 2015

This work integrates chaotic into the recently proposed multilayer neural network algorithm called deep learning. The effects of different chaotic maps on improving the performance of deep learning are investigated. The results demonstrate that chaotic maps are able to improve the performance of deep learning. ©, 2015, Binary Information Press. All right reserved.


Chen G.,Guangdong University of Education | Chen Q.,Guangdong University of Education | Zhang D.,Sun Yat Sen University
Journal of Information and Computational Science | Year: 2013

The purpose of image steganalysis is to detect the presence of hidden messages in cover images. Steganalysis can be considered as a pattern recognition process to decide which class a test image belongs to: the innocent photographic image or the stego-image. We compare harmony search algorithm and particle swarm optimization algorithm based feature selection for image steganalysis. Experiment results show that the proposed hybrid algorithm for feature selection is capable of increasing the testing accuracy of classifying result. The combination of the feature sets extracted with the proposed method is feasible to improve the performance of general steganalysis in a reduced dimension. Experiment results also show that this method has the potential to distinguish different kinds of steganography with the extracted uncorrelated features which contain more discriminatory information. © 2013 Binary Information Press.


Tse C.S.,Chinese University of Hong Kong | Chang J.F.,Guangdong University of Education | Fung A.W.T.,Chinese University of Hong Kong | Lam L.C.W.,Chinese University of Hong Kong | And 3 more authors.
International Psychogeriatrics | Year: 2015

Background: With the proportion of older adults in Hong Kong projected to double in size in the next 30 years, it is important to develop measures for detecting individuals in the earliest stage of Alzheimer's disease (AD, 0.5 in Clinical Dementia Rating, CDR). We tested the utility of a non-verbal prospective memory task (PM, ability to remember what one has to do when a specific event occurs in the future) as an early marker for AD in Hong Kong Chinese. Methods: A large community dwelling sample of older adults who are healthy controls (CDR 0, N = 125), in the earliest stage of AD (CDR 0.5, N = 125), or with mild AD (CDR 1, N = 30) participated in this study. Their reaction time/accuracy data were analyzed by mixed-factor analyses of variance to compare the performance of the three CDR groups. Logistic regression analyses were performed to test the discriminative power of these measures for CDR 0 versus 0.5 participants. Results: Prospective memory performance declined as a function of AD severity: CDR 0 > CDR 0.5 > CDR 1, suggesting the effects of early-stage AD and AD progression on PM. After partialling out the variance explained by psychometric measures (e.g., ADAS-Cog), reaction time/accuracy measures that reflected the PM still significantly discriminated between CDR 0 versus 0.5 participants in most of the cases. Conclusion: The effectiveness of PM measures in discriminating individuals in the earliest stage of AD from healthy older adults suggests that these measures should be further developed as tools for early-stage AD discrimination. Copyright © International Psychogeriatric Association 2014.


PubMed | Chinese University of Hong Kong, Washington University in St. Louis and Guangdong University of Education
Type: | Journal: International psychogeriatrics | Year: 2014

ABSTRACT Background: With the proportion of older adults in Hong Kong projected to double in size in the next 30 years, it is important to develop measures for detecting individuals in the earliest stage of Alzheimers disease (AD, 0.5 in Clinical Dementia Rating, CDR). We tested the utility of a non-verbal prospective memory task (PM, ability to remember what one has to do when a specific event occurs in the future) as an early marker for AD in Hong Kong Chinese. Methods: A large community dwelling sample of older adults who are healthy controls (CDR 0, N = 125), in the earliest stage of AD (CDR 0.5, N = 125), or with mild AD (CDR 1, N = 30) participated in this study. Their reaction time/accuracy data were analyzed by mixed-factor analyses of variance to compare the performance of the three CDR groups. Logistic regression analyses were performed to test the discriminative power of these measures for CDR 0 versus 0.5 participants. Results: Prospective memory performance declined as a function of AD severity: CDR 0 > CDR 0.5 > CDR 1, suggesting the effects of early-stage AD and AD progression on PM. After partialling out the variance explained by psychometric measures (e.g., ADAS-Cog), reaction time/accuracy measures that reflected the PM still significantly discriminated between CDR 0 versus 0.5 participants in most of the cases. Conclusion: The effectiveness of PM measures in discriminating individuals in the earliest stage of AD from healthy older adults suggests that these measures should be further developed as tools for early-stage AD discrimination.


Xu Q.,South China Normal University | Wu J.,South China Normal University | Chen Q.,Guangdong University of Education
Mathematical Problems in Engineering | Year: 2014

Personalized recommended method is widely used to recommend commodities for target customers in e-commerce sector. The core idea of merchandise personalized recommendation can be applied to financial field, which can also achieve stock personalized recommendation. This paper proposes a new recommended method using collaborative filtering based on user fuzzy clustering and predicts the trend of those stocks based on money flow. We use M/G/1 queue system with multiple vacations and server close-down time to measure practical money flow. Based on the indicated results of money flow, we can select the more valued stock to recommend to investors. The experimental results show that the proposed method provides investors with reliable practical investment guidance and receiving more returns. © 2014 Qingzhen Xu et al.

Loading Guangdong University of Education collaborators
Loading Guangdong University of Education collaborators