Dissertation > Excellent graduate degree dissertation topics show

Algorithmic Research on 3D Human Rigid Skeleton Extraction Based on Video Sequence

Author: XuZuoXin
Tutor: WuXiaoJun
School: Harbin Institute of Technology
Course: Control Science and Engineering
Keywords: Computer Vision Motion Capture Skeleton Extraction Rigid body skeleton
CLC: TP391.41
Type: Master's thesis
Year: 2010
Downloads: 109
Quote: 0
Read: Download Dissertation

Abstract


Human motion capture a wide range of applications, including intelligent monitoring, human-computer interaction, virtual reality, three-dimensional computer animation, sports training aids and medical gait analysis. Therefore, human motion capture has a very important significance. The traditional mark points motion capture technology needs such as an LED mark point marked on the body, and then the location and direction of movement of the marked points is sent by the sensor back to the computer and analyzed by computer to obtain the attitude information of the body's movement. Such a motion capture system has the following disadvantages: (1) motion capture equipment is very expensive; (2) dimension was operating personnel need to be familiar with the description and operation; (3) during exercise appear obscured objects and control points, marked exact match is very difficult; (4) of the marked point accurate location information into rigid body motion data is quite difficult. No marker the point motion capture system is usually one or more camera shoot of the characters in the scene, to obtain a three-dimensional model of the human body from synchronized video, and ultimately restore the body's movement, such as joint point position and joint rotation angle. Relative labeled motion capture, without marking motion capture study due to direct the movement of the human body in the video sequence, therefore more research value. With the rigid skeleton of the center of the joint position information to characterize the motion attitude of the body a certain moment, has a very important meaning in the unmarked points motion capture. This study focus on marker motion capture system of three-dimensional rigid body human skeleton extraction algorithm. Current extraction algorithm is sensitive to boundary noise model can not be satisfied with the curve skeleton, and thus unable to calculate the position of the center of the articulation point. To solve this problem, an extract of human 3D skeleton algorithm. First contour image of the video sequence to reconstruct the volume model of the human body; then the repulsive force field theory to calculate body layered curve skeleton; proposed connection guidelines, backtracking on a different limb of the curve skeleton were connected, the connection process using the binary search method to obtain accurate distance and angle threshold, the split and in addition to the the curve skeleton noise; Finally, according to the model of a priori information, to extract the center position of the key points of the human skeleton. The experiments show that the algorithm can get good results with strong robustness to overcome the problem of voxel-based model skeleton extraction algorithm of the model boundary noise sensitive and can achieve automatic extraction of the human skeleton.

Related Dissertations

  1. Research on Learning-Based Low-Level Vision Problem,TP391.41
  2. The Research of 3D Human Motion Capture Based on Reference Points,TP391.41
  3. Research on Nondestructive Detection Technology for External Qualities of Papayas Based-on Vision,S667.9
  4. Study on Detection and Grading of ’Jiro’ Persimmon’s External Quality Based on Computer Vision,S665.2
  5. Prediction of Pork Processing Functionalities,TS251.1
  6. Research on Inspection Technology of Dehydrated Garlic Slice Based on Computer Vision,TP391.41
  7. Moving target trajectory analysis based Intelligent Traffic Monitoring System,TP277
  8. The Design and Implementation of On-line Strip Surface Defect Inspection System Based on Computer Vision,TP391.41
  9. Sensor Data Transmission and Processing of Inertia Motion Capture System,TP212
  10. Learning-based human motion synthesis inverse kinematics,TP391.41
  11. Fusion depth map and 3D models of human motion capture technology research,TP391.41
  12. MEMS-based sensor motion capture system development and design,TP212.9
  13. Cotton fiber maturity longitudinal image analysis,TP391.41
  14. Real-time modeling spatial augmented reality technology research related,TP391.41
  15. Based on sparse representation of human motion capture data analysis method,TP391.41
  16. Application Kinect augmented reality and gesture recognition educational support system,TP391.41
  17. Immersive Virtual Shooting Training System in Human-Computer Interaction Research,TP11
  18. Time series visualization of human motion and multi- index method,TP391.41
  19. Research on Key Technologies of Monocular Visual Odometry Based on DSP,TP391.41
  20. The Technique of 3-dimensional Human Motion Retargeting,TP391.41
  21. Research on Nitrogen Detective System Based on Image Processing,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