Dissertation > Excellent graduate degree dissertation topics show

Research on Face Detection and Tracking Based on Intelligent Visual Surveillance

Author: GaoChunJuan
Tutor: LiuXiangQian；MiaoZhenJiang
School: Beijing Jiaotong University
Course: Signal and Information Processing
Keywords: Face Detection Track MCT AdaBoost Particle filter
CLC: TP391.41
Type: Master's thesis
Year: 2011
Downloads: 147
Quote: 1
Read: Download Dissertation

Abstract


In recent years, the development direction of the video surveillance digitization and network development, is now gradually to the intelligent development. The video surveillance belongs to the category of computer vision. Computer Vision is by computer to mimic the human visual system, so as to achieve the observed world and the purpose of understanding the world. The video human face detection and tracking is a core issue in the field of computer vision. With the development of computer technology and high-speed data storage technology, face detection and tracking with more widely used value and significance, and become one of the hot field of pattern recognition. Face detection and tracking of major value video conferencing, human-computer interaction, video surveillance and access control systems; research value is mainly reflected in the actual environment gives face detection and tracking higher challenge illumination and pose The changes are two problems in the face detection and tracking. Face tracking problem for people in the face detection and video sequences in the results and the latest progress of research at home and abroad on the basis of the research work, including: (1) the face of MCT-based characterization methods, namely the use of texture features of the local structure characterization face mode, the method can be used to resolve the changes in lighting conditions for face detection, the experimental results show that, with the characteristics of the face image in a different light, the light robustness. (2) The combination of (1) a new face characterization methods, AdaBoost algorithm based on local texture features, and the algorithm used for training the classifier, and face detection. Each boosting loop generating a weak classifier in the training of the classifier process of the weak classifiers by a length of 511 index. Adoption of the weighted weak classifiers and the obtained element classification, and finally by the elements classifier, and obtaining a strong classifier. The experimental results show that, compared with the method of Viola, the method can be used a fewer number of cascades, considerable Viola method of face detection results. (3) combined in order to achieve a new face tracking method using particle filter and weighted color histogram. Weighted color histogram considering the relationship between the position of the local distribution of the colors in the image and color space where the. The theoretical framework of particle filter, from the three aspects of the target motion model, modeling of the visual characteristics of the target observation model gives a concrete realization of the target tracking algorithm based on particle filter, a good track.

Related Dissertations

  1. Research on Cooperative Orbit Determination in Satellite Network Based on Multi-Agent System Theory,V474
  2. The Analysis on Characteris of Vibration Source about Vertical Vibration of Surrounding Environment Induced by Rail Transit,U211.3
  3. The Maneuvering Target Tracking Research Based on VRPF,TN957.52
  4. Research on Multi-Sensor Embattling and Continious Tracking Technique Applying for Anti-Stealth and Anti-Interference,TN953
  5. Research on the Classification Based on the Reconstruction of Solder Joint,TP391.41
  6. Tongue Feature Extraction and Research of Fusion Classification,TP391.41
  7. The Fatigue State Recognition of the Driver Based on Eye Detection,TP391.41
  8. Feature Extraction, Selection and Combination in Lipreading,TP391.41
  9. Research on Query Expansion Technique of Retrieval System in Biomedical Field,TP391.3
  10. The Research of 3D Human Motion Capture Based on Reference Points,TP391.41
  11. The Research on Paper Currency Classification Method Based on Harr-Like Feature and Minimal Ball Including Samples,TP391.41
  12. Research on the Theory of Visual Object Tracking Based on Particle Filter for Autonomous Robot,TP242
  13. Research on Predicting Intrinsic Disorder Protein Structure Based on Supervision Manifold Learning Algorithm,Q51
  14. Space Infrared Target Simulation and Application of Target Trace,TP391.41
  15. Face Recognition Method Based on DE,TP391.41
  16. Research on Face Detection Based on Skin Color Segmentation and AdaBoost Algorithm,TP391.41
  17. Detection and Tracking of Moving Object in Complex Background,TP391.41
  18. Analysis and Research on the Cultivation Actuality of Reserve Forces of Heilongjiang Province Curling after the Olympic Winter Games,G82
  19. The Research on the Pattem of the Sports Festival in the Universities in Xi’an,G807.4
  20. Research on Face Recognition Based on AdaBoost Algorithm,TP391.41
  21. The Research on the Target Localization and Tracking Based on WSN,TN929.5

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