Dissertation > Excellent graduate degree dissertation topics show

Head Pose Estimation Based on Active Shape Model

Author: ZhangLei
Tutor: ChenJianXun；Jiang
School: Wuhan University of Science and Technology
Course: Applied Computer Technology
Keywords: Head pose estimation Active shape model Relevance vector machine
CLC: TP391.41
Type: Master's thesis
Year: 2011
Downloads: 94
Quote: 0
Read: Download Dissertation

Abstract


In the field of intelligent surveillance, head pose estimation has great application value. Head pose generally embodies the objects of the individual interest and behavior trends. It’s an important basis of behavior recognition. However, once given a face image, it’s difficult to estimate head pose, especially the pitch and yaw angle. Different faces of different people, the scales changing which caused by the camera position changing, people’s facial expressions changing, the illumination changing and other factors will affect the accuracy of the algorithm.Head pose estimation can be divided into two categories: the image-based and the model-based approaches. The image-based approaches use algorithms such as Principal Component Analysis or Manifold Learning technologies to reduce the dimensions of facial image data, and classify different poses in the low dimensional space. The model-based approaches firstly extract meaningful shape and texture features in the human facial images, which are usually expressed by the parameters model, and then estimate pose based on these characteristics. Opposed to the former, the model-based approaches can capture the subtle head pose change better.This thesis proposes a head pose estimation algorithm based on active shape model. Active shape model is built based on a large number of training shapes. It uses the model parameters which trained by geometric features and texture features to estimate the feature point of the objectives. Active shape model can accurately express the changes of the face shapes of different expressions, different people and different viewpoints. Firstly, the algorithm uses active shape model to locate the facial feature points. Then, predict head pose based on the estimated feature point by relevance vector machine. The experiments on the CAS-PEAL-R1 face database show that the pose estimation by the proposed algorithm is more accurate than the pose estimation by the method of image-based.

Related Dissertations

  1. The image -based face recognition system design and implementation beautification,TP391.41
  2. Feature points based virtual try positioning research,TP391.41
  3. Research on the Technology of Age Estimation Based on Gender Identification,TP391.41
  4. Based LPP video image head pose estimation method,TP391.41
  5. Independent head pose estimation based on the identity of the manifold learning,TP391.41
  6. Research on Facial Feature Points Location Based on Video Images,TP391.41
  7. Robust Head Pose Estimation Algorithm for Embedding Video System,TP391.41
  8. Study on Gaze Tracking in Video Image Sequences,TP391.41
  9. Research and Implementation based on pseudo - muscle vector 3D Facial Animation and its driver,TP391.41
  10. Method of facial features localization,TP391.41
  11. Multi- Manifold Data Modeling and Applications,TP311.13
  12. Research on Some Key Issues of Bayesian Learning Framework Based Nonlinear Manufacturing Process Modeling and Multiobjective Optimization,TP181
  13. Deformation model of technology and its application in medical image segmentation,R318
  14. Study on Human Body Detection and Tracking in Intelligence Visual Surveillance,TP391.41
  15. Gabor Active Shape Model and Its’ Application to Face Segmentation,TP391.41
  16. Research of 3D Reconstruction Key Technology Based on Monocular Video,TP391.41
  17. Automatic Human Age Estimation Based on Layered and Weighted Support Vector Machine Model,TP391.41
  18. Face Recognition feature extraction algorithm,TP391.41
  19. Emotional geometrical features and support vector machine - based face expression recognition research,TP391.41
  20. Research on Technologies of Semantic-Based Medical Image Retrieval,TP391.41
  21. Study on the Morphing of Color Facial Images Based on Improved MR-ASM,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