Dissertation > Excellent graduate degree dissertation topics show

Research on Human Identification Based on Gait Analysis

Author: HouBenBo
Tutor: WangKeJun
School: Harbin Engineering University
Course: Pattern Recognition and Intelligent Systems
Keywords: Gait Recognition Gait energy image (GEI) Cycle detection Linear discriminant analysis (LDA) K Nearest Neighbor (KNN)
CLC: TP391.4
Type: Master's thesis
Year: 2007
Downloads: 170
Quote: 4
Read: Download Dissertation


Gait recognition in recent years, the field of computer vision and biometrics research direction of a concern, it is aimed at people walking posture identification. Compared to other biometric identification technology, gait recognition advantage lies in the non-contact, non-invasive, easy to perceive, difficult to hide, difficult to disguise. Based on the above characteristics, gait recognition has broad application prospects and economic value in areas such as access control systems, security and surveillance, human-computer interaction, medical diagnosis, and therefore stimulate the enthusiasm of the majority of domestic and foreign research scientists. Gait recognition mainly consists of four parts: the gait image preprocessing, cycle detection, feature extraction and gait recognition. Feature extraction is a key factor of the gait recognition, and therefore the focus of this study. Gait recognition, in-depth research, mainly in the following areas of work: Research gait recognition, and combine literature gait recognition factors, performance prediction, a more comprehensive elaborated, followed by a synoptic overview of the four components of gait recognition. Analysis of the motion detection algorithm, and used according to the specific circumstances of the background subtraction algorithm to achieve the detection of the movement of the human body, morphological operations and image geometry transform image standardization and centralization, and a whole new cycle detection The algorithm used to achieve the cycle division, and in accordance with the periodic sequence image generation is used to describe the gait characteristics of gait energy image. Feature extraction method gait recognition at home and abroad, mainly divided into two categories: model-based approach and non-model method. Need to create a model-based method of the body structure model or body motion model, then Get characterized; non-model method do not need to create these models, but to create a corresponding relationship between the periodic sequence based on characteristics such as position, velocity, shape information. Often encounter the problem of linear discriminant analysis (LDA), that due to the small sample sample within-class scatter matrix is ??usually singular, and improved LDA algorithm to seek the optimal solution of characteristic zero space to achieve step state of low-dimensional feature extraction. Based KNN classifier and the Euclidean distance metric based on the extraction of low-dimensional feature match of the sample in the test sample and the sample library, gait-based identification. The library of the Chinese Academy of automated gait, the number of individuals in the sample library of 124 people, side Perspective recognition rate of 98.7%. Looking further research and ideas on the basis of this study: a multi-angle gait feature fusion, 3D modeling and multi-camera use, data fusion, and based on the complex context of large-scale gait database performance evaluation.

Related Dissertations

  1. Signal Processing Method Based on Singular Value Decomposition and Its Application to Mechanical Fault Diagnosis,TH165.3
  2. Identity -based human gait recognition algorithm based on distance,TP391.41
  3. Research and Implementation of the Key Technologies in Gait Recognition,TP391.41
  4. Gait Recognition Research and Implement Based on Image Transformation,TP391.41
  5. Gait Recognition Based on Joint Points,TP391.41
  6. OpenCV Based on Gait Recognition System,TP391.41
  7. Door gates Control System,U231.6
  8. Gait-based identification technology research,TP391.41
  9. KNNModel algorithm and its application,TP311.13
  10. Research on Gait Object Extraction and Recognition,TP391.41
  11. The Study of Gait Recognition Based on Image Sequence and Pressure,TP391.41
  12. Personal Identification Based on Two Dimensional Gait,TP391.41
  13. Computer face recognition technology and applied research,TP391.41
  14. Based on dual-source multi-feature fusion technology research gait,TP391.41
  15. Dimensionality reduction algorithm based on subspace biometric Applied Research,TP391.41
  16. Research on Technology in Identification of Aerial Targets Based on Support Vector Machine,TN953
  17. Theories and Algorithms of Manifold Learning and Applications in Biometric Authentication,TP181
  18. The Research on Canonical View Synthesis and Gait Energy Image for Human Identification,TP391.41
  19. Facial Expression and Gender Recognition Based on Static Facial Image,TP391.41
  20. Angle of gait recognition using statistical characteristics,TP391.41
  21. The Research of Mainstream Feature Extraction and Integration for Face Recognition,TP391.41

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