Dissertation > Excellent graduate degree dissertation topics show

The Research on Facial Feature Points Alignment Technology Based on Shape Model

Author: ShiHongWei
Tutor: LiLiJuan
School: Hunan University
Course: Applied Computer Technology
Keywords: Face Recognition Face Detection Feature location Shape model Local Binary Pattern Weighted Bayesian tangent model
CLC: TP391.41
Type: Master's thesis
Year: 2009
Downloads: 98
Quote: 1
Read: Download Dissertation

Abstract


Automatic Face Recognition (AFR) technology the ability to try to give the computer based on the facial feature to distinguish human identity, the study has important scientific significance and great value. After 30 years of development, AFR technology has made considerable progress. Currently, the best face recognition systems under ideal conditions have been achieved acceptable recognition performance, testing and practical experience shows that: under non-ideal conditions, face recognition technology is far from mature, to develop a truly robust, The practical application of AFR systems also need to address the key issues, especially the need to study the precise positioning of facial key features as the identification of the necessary preconditions. Feature location is far from being a problem already solved, in particular, need to pay attention to the the feature location under non-ideal imaging conditions. Therefore, in order to develop a real practical face recognition systems, we must pay full attention to the positioning of key feature points on the face, the main work of this paper: This paper focuses on the problem of facial features based on statistical learning point positioning details ASM basic principles expounded of of / BTSM model-based grayscale appearance model the ASM / BTSM algorithm, the system introduced the ASM / BTSM model extension of the ASM / BTSM deformation mode, iterative search grayscale appearance match process The specific depth analysis. A shape model positioning algorithm based on local binary pattern. The algorithm extracts feature points to the center of the rectangular area LBP block as the local features of facial feature points, using the principal component analysis algorithm to build a distribution model based on the reconstruction error, and then treat the match the shape of the iterative search. Further, in order to reduce the error caused on the pending matching point is not in the normal direction, using the eight-direction search the optimum contour point. The experiment compares the proposed algorithm with traditional algorithm has the advantages, LBP-BTSM algorithm is more accurate than traditional feature point location algorithm, but the attitude problems still affect the positioning of the characteristics of the main problems. Proposed a shape evaluation based on the weighted the the Bayesian the tangents model (WBTSM). On the basis of the local texture model based define a shape of an evaluation function, which measures the degree of matching of the search to get the shape of the training data. WBTSM shape evaluation information, to search for the shape obtained by weighted manner projected into shape subspace, and unlike in the Bayesian shape model using orthogonal projection, compared with the orthogonal projection, the weighted projection can use the search process information, the search might jump out of local optima, so as to get more accurate results. Finally, we designed a windows-based automatic face recognition system platform can effectively extract facial features, the system includes face detection feature location, face recognition three main components.

Related Dissertations

  1. Research on Algorithms of 2D Face Template Protection,TP391.41
  2. Study on Model-Based Visualization of Wheat Root Systems,S512.1
  3. Research of Video Face Recognition Based on Weighted Voting and Key-Frame Extraction,TP391.41
  4. Face Recognition Method Based on DE,TP391.41
  5. Research on Face Detection Based on Skin Color Segmentation and AdaBoost Algorithm,TP391.41
  6. The Research of Image Matching Method Based on Feature Descriptor,TP391.41
  7. Research on Face Recognition Methods Based on Flexible Neural Tree,TP391.41
  8. Research on Face Recognition Based on AdaBoost Algorithm,TP391.41
  9. Research Onamethod for Human Face Recognition Based on MMTD,TP391.41
  10. The Research on Neural Network with Quadratic Denominator Cubic Rational Spline Function Weight and Its Application,TP183
  11. The image -based face recognition system design and implementation beautification,TP391.41
  12. Face recognition algorithm based on feature fusion research,TP391.41
  13. The Design of Identification System for Power Marketing Services,TP391.41
  14. Based on wavelet transform and linear subspace face recognition technology,TP391.41
  15. \,R338
  16. Based on local binary pattern and Weber local descriptor Face Recognition,TP391.41
  17. Expression of face recognition algorithm based on sparse research,TP391.41
  18. Color Face Detection and Recognition,TP391.41
  19. Under the conditions of variable lighting and occlusion of face recognition technology and its application,TP391.41
  20. Based on sparse representation Face Image Recognition Method,TP391.41
  21. Research on Subspace Analysis Based Face Recognition Algorithm,TP391.41

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