Dissertation > Excellent graduate degree dissertation topics show

Parallelization Research of Face Recognition and Texture Synthesis Algorithm Base on MATLAB Multi-core Clusters

Author: YuMengLing
Tutor: ZhengXiaoZuo
School: Liaoning Normal University
Course: Computer System Architecture
Keywords: MATLAB multi-core clusters parallel computing face recognition algorithmof PCA Image Quilting texture synthesis algorithm task partition parallel strategies
CLC: TP391.41
Type: Master's thesis
Year: 2012
Downloads: 5
Quote: 0
Read: Download Dissertation

Abstract


In order to take reasonable advantage of multi-core clusters and improve the available efficiency of CPU, parallel optimization was studied which according to the algorithm of PCA-based face recognition and Image Quilting texture synthesis based on MATLAB distributed parallel computing cluster. Deeply analyzing the cluster architecture, the hardware and software system structure, algorithm programming model and so on, and combined with structure features of serial program complete the parallel programming process, which can play a well potential capacity of multi-core and promote the performance of multi-core application.The research focus is as follows:For one thing, according to the algorithm of PCA-based face recognition, parallel programming model by building blocks with multithreading was studied to improve the performance of the program. According to the integral structure of PCA-based face recognition algorithm, a functional module named train() was designed for the training of recognizing generated samples in the environment of MATLAB cluster. The parallelization of this algorithm was realized by task partition. The experimental results indicate that the stable recognition rate of94.167%and the approximately linear speedup ratios verify the correctness and the high efficiency of the algorithm parallelization, and the development of parallel application was condensed further. Parallel method caused little influence of the algorithm through the MATLAB distributed parallel computing cluster, and better solved intensive computing optimization problems of large amount of data in multi-core computer or multi-core clusters. For another thing, according to Image Quilting for texture synthesis algorithm characteristic that the new and old texture pieces spliced, a parallel programming model that creating multi-task realized the texture pieces to cut and splice was studied to Image Quilting for texture synthesis algorithm. A functional module named iq() was designed for the calculation of suture path in the overlapping area, subtasks assigned to each Worker executed parallel computing by using of the scheduler of MATLAB multi-core clusters. The experimental results indicated that this algorithm can achieve a better speed-up ratio and enhance the available efficiency of multi-core CPU and promote the utilization rate of multi-core computer resources effectively.The characteristics of the research was to design the MATLAB parallel applications through multi-thread creating, dynamic configuration task manager, thread number setting and thread scheduling strategy in order to improve the algorithm performance. In addition, the attention of algorithm’s improvement and development should be paid to data sharing, module partition, task parallelism and distribution and load balancing between processor core numbers, which to ensure each nucleus of the multi-core processors in a high load condition and accomplish parallel computing task accurately.It effectively modified the bottleneck of serial program in the algorithm of PCA-based face recognition and Image Quilting texture synthesis through theverification of MATLAB multi-core cluster platform, which have improved the working efficiency of CPU and made full use of multiple nuclear resources. At the same time, it also demonstrated the advantages that algorithm application of graphic and image combining with multi-core parallel computing technology.

Related Dissertations

  1. The Airborne LiDAR and LiDAR Points-cloud’s Quick Processing Method,TN959.73
  2. Study of 3D Stratum Modelling and Visualization Method Based on MPI,TP391.41
  3. The Research of Orthophoto Generation Based on GPU Parallel Acceleration,TP391.41
  4. The Research on Feature Selection for Data Stream,TP311.13
  5. Research of Sub-Diffraction in Optical Diffraction Field,O436.1
  6. Parallel Computing and Performance Analysis of Concurrent Systems,TP338.6
  7. Study on Haze Removal Method and Parallel Implementation for HJ-1 Satellite CCD Image,P228
  8. Research on Medical Ultrasound Imaging Techniques Using Parallel Computing,TP391.41
  9. Research of Finite Element Method on GPU,O241.82
  10. Research on Energy-saving and Emission Reducting Generating Dispatch Based on Multi-Objective Intelligence Algorithm,TM73
  11. The Algorithm Researches of Novel Wide Area Backup Protection for Power Grid,TM774
  12. The Research on Online Adaptive Settings,TM77
  13. GPU-based parallel search algorithm for time series,TP391.41
  14. Research of the Grid Computing and Parallel Technology Based Dynamic Security Assessment of Power Systems,TM712
  15. The Research and Realization on the Spatial Computing Models for Huge Spatial Data,P208
  16. The Study on the UAV Digital Remote Sensing & Survey System Integration and Images Data Processing,P237
  17. A Parallel DDM Preconditioner for Solving Finite Element Discretization of Elasticity Problem in Three Dimension,O241.82
  18. Based on parallel environment for TSP,O224
  19. Lattice-Boltzmann Simulation on Fluid Dynamics and Mixing in Micro Channels,TQ021.1
  20. The Study on the Computing System of Relay Coordination for Extra-High Voltage Power Grid Based on .NET Platform,TM774
  21. The Research and Realization of the Partitioning Strategy of Vector Spatial Data in Parallel Computing Environment,P208

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net  Mobile