Shenzhen Institute of Information Technology

www.sziit.com.cn/
Shenzhen, China

Shenzhen Institute of Information Technology is a university located in Shenzhen, Guangdong, China. Its current president is Zhang Jihong. Wikipedia.

SEARCH FILTERS
Time filter
Source Type

Wei Y.,Shenzhen Institute of Information Technology
RISTI - Revista Iberica de Sistemas e Tecnologias de Informacao | Year: 2016

This paper used the recursive relation among the generating binary trees with n nodes, considered the trees as a type according to the number of nodes of its left and right subtree firstly, then found out the sequences in each type of them, thus achieved the binary trees in lexicographic order. Not only can generate the next tree according to any trees, but find the corresponding binary tree according to the random number. This provided the basis for security mechanism with the encryption/decryption of multi binary trees.


Zheng Y.,Harbin Institute of Technology | Zhao W.,Shenzhen Institute of Information Technology
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2017

Roadside camera calibration is essential to intelligent traffic surveillance and still an unsolved problem. The commonly used pattern-based calibration methods are suitable for the laboratory environment rather than real traffic environment, since the calibration patterns (e.g., checkerboards) generally do not exist in traffic scenarios. In view of this, we propose a new framework for roadside camera calibration where the vehicle moving on the roadway is first introduced as a calibration pattern. Considering that the vehicles are main monitoring targets and inevitably appear in traffic scenarios, the proposed calibration method has a wide use range and is not limited to the structure information of traffic scenarios. Inspired by the traditional pattern-based calibration methods that utilize the matching of 3D-2D point correspondences, we utilize the 3D-2D vehicle matching for camera calibration. The key insight is to convert the camera calibration problem into a vehicle matching problem. To improve the accuracy of calibration results, a new measure function is provided to evaluate the vehicle matching degree and a dynamic calibration method using multi-frame information is proposed to correct camera parameters. Experiments on real traffic images demonstrate the effectiveness and practicability of the proposed calibration framework. © Springer International Publishing AG 2017.


Ma C.,Shenzhen Institute of Information Technology
Journal of Digital Information Management | Year: 2017

Recently extreme learning machine (ELM) was proposed as a new learning method for single hiddenlayer feedforward neural networks (SLFNs), it is not the same as traditional gradient based learning algorithm strategies as it can achieve good generalization performance as well as extremely fast learning speed. However, ELM may require large number of hidden neurons due to the random determination of the input weights and hidden biases, and there may exist a set of non-optimal parameters which lead ELM not be able to reach the global optimum in some cases. With the help of ideas that using a hybrid approach which takes advantage of the optimization method and ELM to train SLFNs, this study proposes a novel hybrid approach based on artificial bee colony (ABC) optimization method to optimize the ELM parameters, where the optimal input weights and biases of ELM are specified by the ABC approach and Moore-Penrose (MP) generalized inverse to analytically determine the output weights. The proposed algorithm, named ABC-ELM, is rigorously compared with the original ELM and other evolutionary ELM methods in different classification datasets. The obtained results clearly confirm that the proposed approach is more suitable for classification problems that we investigated, and it can not only achieve better generalization performance but be more robust with much more compact networks.


Jiang W.,Shenzhen Institute of Information Technology
Proceedings - 2017 16th ACM/IEEE International Conference on Information Processing in Sensor Networks, IPSN 2017 | Year: 2017

In this study, we propose a crowd sensing framework with the existence of execution uncertainty and a given budget. Our framework consists of three stages: Task Selection, Task Allocation, and Payment. Within each stage, we define the design problems and give a preliminary solution with desirable theoretic properties.


Li S.,Stevens Institute of Technology | Chen S.,Shenzhen Institute of Information Technology | Liu B.,University of Massachusetts Amherst
Neural Processing Letters | Year: 2013

Bartels-Stewart algorithm is an effective and widely used method with an O(n 3) time complexity for solving a static Sylvester equation. When applied to time-varying Sylvester equation, the computation burden increases intensively with the decrease of sampling period and cannot satisfy continuous realtime calculation requirements. Gradient-based recurrent neural network are able to solve the time-varying Sylvester equation in real time but there always exists an estimation error. In contrast, the recently proposed Zhang neural network has been proven to converge to the solution of the Sylvester equation ideally when time goes to infinity. However, this neural network with the suggested activation functions never converges to the desired value in finite time, which may limit its applications in realtime processing. To tackle this problem, a sign-bi-power activation function is proposed in this paper to accelerate Zhang neural network to finite-time convergence. The global convergence and finite-time convergence property are proven in theory. The upper bound of the convergence time is derived analytically. Simulations are performed to evaluate the performance of the neural network with the proposed activation function. In addition, the proposed strategy is applied to online calculating the pseudo-inverse of a matrix and nonlinear control of an inverted pendulum system. Both theoretical analysis and numerical simulations validate the effectiveness of proposed activation function. © 2012 Springer Science+Business Media, LLC.


Mu X.,Shenzhen Institute of Information Technology
2010 Chinese Control and Decision Conference, CCDC 2010 | Year: 2010

A fuzzy neural sliding mode controller based on genetic algorithm (FNSMCGA) is presented for trajectory tracking control of multi-link robots with model errors and uncertain disturbances. This approach gives a new global sliding mode manifold for multi-link robots, which enable system trajectory to run on the sliding mode manifold at the start point and eliminate the reaching phase of the conventional sliding mode control. Robustness for system dynamics is guaranteed over all the response time. A fuzzy neural network (FNN) is employed to eliminate chattering of global sliding mode control, and enforce the sliding mode motion by FNN learning the upper bound of model errors and uncertain disturbances. Genetic algorithm can optimize the FNN initial parameters, which can make the robot running with expected trajectory in whole running process. The control laws are calculated by Lyapunov stability method, which ensure that the controlled system is stable. Simulation results verify the validity of the control scheme. ©2010 IEEE.


Zhan S.,Shenzhen Institute of Information Technology | Huo H.,Shenzhen Institute of Information Technology
Journal of Information and Computational Science | Year: 2012

Job scheduling system problem is a core and challenging issue in cloud computing. How to use cloud computing resources efficiently and gain the maximum profits with job scheduling system is one of the cloud computing service providers' ultimate goals. For characteristics of particle swarm optimization algorithm in solving the large-scale combination optimization problem easy to fall into the search speed slowly and partially the most superior, the global fast convergence of simulated annealing algorithm is utilized to combine particle swarm optimization algorithm in each iteration, which enhances the convergence rate and improves the efficiency. This paper proposed the improve particle swarm optimization algorithm in resources scheduling strategy of the cloud computing. Through experiments, the results show that this method can reduce the task average running time, and raises the rate availability of resources. 1548-7741/Copyright © 2012 Binary Information Press.


Huo H.,Shenzhen Institute of Information Technology | Zhan S.,Shenzhen Institute of Information Technology
International Journal of Digital Content Technology and its Applications | Year: 2012

Although the grey forecasting model has been successfully utilized in many fields and demonstrated promising results, there are some problems in GM(1,1) model, such as, model method biased, transformation inconsistent and first number of the initial sequence not functioning high precision prediction in model after an accumulated generating operation. Literatures show its performance still could be improved. For this purpose, this paper proposes an improved grey GM(1,1) model, which uses Fourier series to correct the residual of original value and predictive vale, and reconstructs the GM(1,1) white background value based on genetic algorithm. As shown in simulation results, the proposed model obviously can improve the prediction accuracy of the original grey model, and has a very high practicability and reliability.


Chen Y.-Q.,Shenzhen Institute of Information Technology
International Journal of Theoretical Physics | Year: 2013

One-parameter general coherent state of the gl(2,1) superalgebra is constructed. Its properties are discussed in detail. One-parameter matrix elements of the gl(2,1) generators in the one-parameter general coherent-state space are calculated. © 2012 Springer Science+Business Media New York.


Patent
Shenzhen Institute of Information Technology | Date: 2014-06-04

The invention provides a method and device of extracting a sound source acoustic image body in 3D space. The method includes: determining a spatial position of a sound source acoustic image and determining a speaker beside the spatial position where the sound source acoustic image is located according to the determined spatial position (, , ) of the sound source acoustic image; calculating a correlation of signals of all sound tracks of the selected speaker in the horizontal direction and the vertical direction, and obtaining and storing a parameter set {IC_(H), IC_(v), Min{IC_(H), IC_(v)}} of a acoustic image body, wherein the Min{IC_(H), IC_(v)} is a smaller value between IC_(H )and IC_(v). The expression parameters of the acoustic image body obtained in the present invention are used for providing technical support for accurately restoring the size of the sound source acoustic image in a 3D audio live system, which solves the technical problem that the restored acoustic image in a 3D audio is excessively narrow at present.

Loading Shenzhen Institute of Information Technology collaborators
Loading Shenzhen Institute of Information Technology collaborators