Bui Thu Lam, Le Quy Don Technical University
Deputy Editor-In-Chief
Tran Xuan Nam, Le Quy Don Technical University
Advisory Member
Pham The Long, Le Quy Don Technical University
Nguyen Binh, Posts and Telecommunications Institute of Technology
Dinh The Cuong, Directorate of Information Technology
Luong Chi Mai, Vietnam Academy of Science and Technology
Huynh Quyet Thang, Hanoi University of Science and Technology
Nguyen Huu Thanh, Hanoi University of Science and Technology
Vu Duc Thi, Vietnam National University, Hanoi
Nguyen Thanh Thuy, Vietnam National University, Hanoi
Dao Thanh Tinh, Le Quy Don Technical University
Scientific Secretary
Nguyen Van Giang, Le Quy Don Technical University
Table: Current Issue
ID Paper
[Pdf link]
Automatic Localization and Navigation Assistance for Visually Impaired People using Robot.
Nguyen Quoc Hung, Vu Hai, Tran Thanh Hai, Nguyen Quang Hoan
This paper presents a new support system for the visually impaired navigation in small pervasive environments using mobile robots. The system uses computer vision techniques to build environment map, locate and detect obstacle. The proposed method enables the robot to quickly know its location on the map. The object detection and recognition algorithms are performed in real time in order to avoid static and dynamic obstacles in the environment. To help the visually impaired person navigating in the environment, the proposed method aims to find the optimal path that is the shortest and the most secure path for the moving of the human. In order to interact between human and robot, we design and develop an interface using touch screen and vibration patterns of a smart-phone device. The system is evaluated on a number of blind pupils. The experimental results confirmed that the proposed system is feasible to deploy in practical application.
Application Access Control on Smartphones using Face Recognition.
Cao Tuan Dung, Trinh Thanh Trung
This paper presents a systematic study of creating access controlling feature on smartphones using face recognition. We focused on comparative assessment of different approaches in each basic processing step in recognition process. We proposed a number of improvements in pre-processing techniques and using the combination of LBP in face detection and LBPH in face recognition. The experimental results in the computing environment, which is smartphone with standard and custom data sets, provide a reduction of over 50% in average recognition time compared to solutions using Haar and Eigenfaces algorithms. The highest recognition precision is 85.22% achieved in situations which recognition photos have similar lighting conditions with trained photos.
New NHPP SRM based on Generalised S-shaped Fault-Detection Rate Function.
Nguyen Hung Cuong, Huynh Quyet Thang
Software reliability modelling (SRM) is a mathematics technique to estimate some measures of computer system that relate to software reliability. One of the best approaches is using non-homogeneous Poisson process (NHPP). In this group of models, fault-number and failure-rate are constant or time-dependent functions. A few studies have been manipulated S-shaped curve to construct their models. However, those works remain some limitations: the reality of used functions; the complex computing limits the way to build new SRM. In this study, we introduce a new model that is based on a generalised S-shaped curve and evaluate it by real data set. After installing it in real code of Matlab and using MLE method to estimate parameter with a range of initial solutions, we confirm that our model converge to the most basic model of NHPP group, Goel-Okumoto model.
Locate Resources for Tasks on Cloud Computing Based on QoS Constraints.
Nguyen Hoang Ha, Le Van Son, Nguyen Mau Han
The purpose of the task scheduling is to find an efficient method of mapping between sets of tasks in a subset of resources to achieve a reasonable schedule. The schedule given to ensure tasks completed before Deadline and satisfy budget of the task. This paper gives an algorithm for mapping from the set of tasks with input parameters such as the time, deadline, budget and workload to the subsets of resources with costs and speeds is different. We construct the problem as a constraint optimization problem and give an algorithm with polynomial time complexity to map effectively the resources with target completion costs of the action is smallest but still satisfying deadline and budget of the task. After that we use CloudSim simulation to compare this algorithm with the algorithm Earliest Deadline First (EDF).
Post-processing for Handwriting Recognition using Vietnamese Person Name Dictionary.
Pham Van Hung
This paper presents a model of post-processing for Vietnamese name to increase the accuracy for handwriting in form. The current model is base on the grammar and statics. Thus, they are not good for name person recognition. Because the person name seem is not accordance with rule of grammar at all. If a person name is a word then the number is very large in the millions versus thousands of syllabic units. In addition, the sign of the Vietnamese uppercase handwriting letters is small to recognizer. Our model solves the problem above. Specifically, the accuracy of Vietnamese person name handwriting recognition is 96.03% using person name dictionary compare to 77.81% using syllable dictionary. Our approach is embedded into the automatic data entry system for enrollment forms, VnHandwritten1.0, the first commercial application in recognizing Vietnamese handwriting in Vietnam. On this application, our method proves itself an efficient method for commercial purpose.
Add-Border-SMOTE: A New Over-sampling Method for Imbalanced Data Sets.
Nguyen Thi Hong, Nguyen Manh Cuong, Dang Xuan Tho
This paper introduces about imbalanced data sets, the importance of imbalanced data sets classification and analyses some approach methods. Then, we presents a new method over-sampling for minority class Add-Border-SMOTE. Add-Border-SMOTE is based on Borderline-SMOTE method to find out minority examples on borderline or near borderline and generate synthetic examples between borderline examples.
VLSI Architecture of Magnitude Estimation Algorithm for Speech Recognition System.
Nguyen Ly Thien Truong, Hoang Trang
In speech recognition system, speech signal is preprocessed, and then will be transferred to the frequency domain by Fast Fourier Transform (FFT). The output of FFT block is the complex numbers that will be estimated in magnitude to prepare for the calculation of Mel-frequency filter bank block. This paper proposes a novel “alpha max-beta min” algorithm for magnitude estimation whose hardware implementation is very simple. The time executes proposed algorithm in only one clock cycle, and average error of less than 1%. The influences of proposed algorithm on the whole speech recognition system are also given in calculation speed and recognition accuracy.
The Blind Signature Scheme Based On Square Root Problem.
Nguyen Tien Giang, Nguyen Vinh Thai, Luu Hong Dung
This paper proposes a blind signature schema developed from a digital signature scheme based on the difficulty of the square root probem on the polynomial flange $Z_{n=p.q}$, where p, q are distinguished primes. The new signature scheme proposed has higher safety level compared to the schemas which have been published previously about the ability of keeping secret the source of the signed messages.
An Alternating Optimization Method to Improve Sum Rate of MIMO Interference Channels.
Ha Hoang Kha
The channel capacity of interference channels has been an active research area in recent years. This paper is concerned with the multi-user wireless networks where multiple users are simultaneously sharing the same frequency spectrum. Each user is equipped with multiple antennas to beam the transmitted signals to desired directions, and to linearly combine the received signals in order to mitigate interference. Two typical methods to design transceiver matrices are the minimization of leakage interference and the maximization of stream signal to interference plus noise. However, previous methods commonly assume that the power allocated to all signal streams is equal. The current paper will optimize the power allocation to streams to improve the sum-rate. In addition, the evaluation of the sum rate performance, the impacts of initializations and convergence speed have not been extensively studied in the literature. By using Monte-Carlo simulation, the current paper will investigate the affects of these parameters to the system performance.
A Case for Multipath-based Multicasts on Random Topologies.
Ngo Quang Vinh, Hoang Trang, Vu Dinh Thanh, Ikki Fujiwara, Michihiro Koibuchi
Network latency is a performance concern especially for collective communication because communication overhead for small messages dominates the system performance in both off-chip and on-chip large-scale interconnection networks. The aggregate path hops strongly affect the network latency in collective communication. In this context we try to minimize the number of total packet path hops for collective communication by introducing a combination of path-based multicastings, especially for random topologies. From our graph analysis results, multicasts based on multiple paths that include all the destinations can reduce the aggregate path hops by up to 5.5\% when compared to path-based if the number of destinations is small. Our recommendation to support multicasts for a few destinations is to use multicasting based on multiple paths.